U.S. Government Restricted/Limited Rights - Esri Supportdownloads2.esri.com/.../ao_/Modeling_our_World.pdf · Modeling Our World The ESRI Guide to Geodatabase Design ISBN 1-879102-62-5

Copyright © 1999 Environmental Systems Research Institute, Inc.All r ights reser ved. Printed in the United States of America.

The information contained in this document is the exclusiveproper ty of Environmental Systems Research Institute, Inc. Thiswork is protected under United States copyright law and thecopyright laws of the given countr ies of origin and applicableinternational laws, treaties, and/or conventions. No par t of thiswork may be reproduced or transmitted in any form or by anymeans, electronic or mechanical, including photocopying orrecording, or by any information storage or retrieval system,except as expressly permitted in writing by EnvironmentalSystems Research Institute, Inc. All requests should be sent tothe attention of Contracts Manager, Environmental SystemsResearch Institute, Inc., 380 New York Street, Redlands,California 92373-8100 USA.

The information contained in this document is subject to changewithout notice.

U.S. Government Restricted/Limited Rights

Any software, documentation, and/or data deliveredhereunder is subject to the terms of the LicenseAgreement. In no event shall the U.S. Government acquiregreater than RESTRICTED/LIMITED RIGHTS. At aminimum, use, duplication, or disclosure by the U.S.Government is subject to restr ictions as set for th in FAR§52.227-14 Alternates I, II , and III ( JUN 1987); FAR §52.227-19 ( JUN 1987) and/or FAR §12.211/12.212 (CommercialTechnical Data/Computer Software); and DFARS §252.227-7015 (NOV 1995) (Technical Data) and/or DFARS§227.7202 (Computer Software), as applicable. Contractor/Manufacturer is Environmental Systems Research Institute, Inc.,380 New York Street, Redlands, California 92373-8100 USA.

PUBLISHED BY

Environmental Systems Research Institute, Inc.380 New York Street

Redlands, California 92373-8100

ESRI, MapObjects, ARC/INFO, and ArcView are trademarks ofEnvironmental Systems Research Institute, Inc., registered in theUnited States and certain other countries; registration is pending inthe European Community. ArcInfo, ArcMap, ArcCatalog, ArcObjects,AML, ArcSDE, ArcIMS, ARC GRID, Arc Explorer, and the ESRI Presslogo are trademarks and www.esri.com is a service mark ofEnvironmental Systems Research Institute, Inc.

The names of other companies and products mentioned herein aretrademarks or registered trademarks of their respective trademarkowners.

Environmental Systems Research Institute, Inc.Modeling Our World

The ESRI Guide to Geodatabase DesignISBN 1-879102-62-5

Preface

All geographic information systems (GIS) are builtusing formal models that describe how things arelocated in space. A formal model is an abstract andwell-defined system of concepts. It defines thevocabulary that we can use to describe and reasonabout things. A geographic data model defines thevocabulary for describing and reasoning about thethings that are located on the earth. Geographic datamodels serve as the foundation on which allgeographic information systems are built.

We are all familiar with one model for geographicinformation—the map. A map is a scale model ofreality that we build, using a set of conventions andrules (for example, map projections, line symbols,text). Once we construct a map, we can use it toanswer questions about the reality it represents. Forexample, how far is it from Los Angeles to SanDiego? Or, what cities lie along the Mississippi River?The map model also serves as a tool forcommunicating facts about geography visually: Is theterrain rough? Which way is north? In fact, when wesee a map, we often understand things that might noteven occur to us as specific questions.

Maps work because we know the “rules” ofconventional map reading: blue lines are rivers,North is toward the top of the page, and so on. In asimilar way, geographic data models define theirown set of concepts and relationships, which mustbe understood before you can expect to create orinterpret your own data model. These conceptsrelate to how you can represent geographicinformation in a computer system, rather than, as inthe map example, on paper.

In Modeling Our World, Michael Zeiler has writtenan excellent primer for understanding the variousmodels used to represent geographic information inArcInfo™ 8 software. He presents, usingstraightforward text and excellent illustrations, theconcepts and vocabulary employed in the design,implementation, and use of the ArcInfo 8 geographicdatabase. In addition to explaining the ArcInfo data

model (objects, features, surfaces, networks, images,and so forth) in detail, Michael also provides goodinsight into how to use this framework to designuseful information models that fit your particularneeds.

This book serves a variety of different purposes. Forthe geographer or scientist, it defines a conceptualcontext for representing geographic information. Forthe GIS specialist, it serves as a guidebook indesigning and using geographic databases. Finally, itintroduces database concepts to a geographicaudience, and geographic concepts to the databasespecialist.

ArcInfo 8 defines a unified framework forrepresenting geographic information in a database.Several different generic data models are supportedwithin this framework:

• cell-based or raster representation

• object-based or feature-based representation

• network or graph-element representation

• finite-element or TIN representation

Each of these generic models has its own vocabularyused to define and reason about geographicinformation. When we decide to represent roads,rivers, terrain, or any sort of phenomena in a GIS,we need to decide exactly how we defineinformation in terms of these generic models. Aschapter 1 points out, there are many ways thatinformation can be modeled in a GIS. Therepresentation you choose for the data model willaffect how you sample and measure geographicinformation, how you display it visually, and whichrelationships between elements can be represented,as well as query and analysis operations that can beapplied to the information.

Some have asserted that we should hiderepresentational models for geographic information(features, geometry, rasters, surfaces, and so on)

from the users of geographic information systems.Somehow, these representational concepts areconsidered “implementation details.” In this view, asingle real-world thing, such as the Mississippi River,should be modeled as a single thing within the GIS.Perhaps, behind the scenes, the system couldautomatically use multiple representations for thesereal-world things. If you ask “What is upstream?” itcould use a network representation of the river. Ifyou ask “What is the surface area of the water?” itcould use a polygon feature representation. If youask “What area does it drain?” it could use a surfaceor terrain representation, and so on. While it may bedesirable to hide these concepts from someconsumers of geographic information, I believe thata strong understanding of geographic data modelsand representations is crucial to the correct designand use of geographic information systems.Geographic data models act as the lens or filterthrough which we perceive and interpret the infinitecomplexity of the real world. It is only in the contextof representations of the Mississippi River, forexample, that we can define specific properties,behavior, or even its identity as a “thing of interest.”Understanding geographic data model concepts iscentral to knowing how to define and collectgeographic information. It is also crucial for correctlyinterpreting the results derived from the analysis ofgeographic information. This is similar to the rolethat statistics and sampling theory play in the naturalsciences.

For the GIS specialist, this book serves as anintroduction to a new object-relational model forrepresenting features, spatial relationships betweenfeatures, and other thematic relationships. This newmodel is significantly richer in its ability to representfeatures with associated behavior, relationships, andproperties than the current coverage or shapefilemodel. If you are already familiar with coverages,shapefiles, and database tables, the new model is adramatic extension of concepts and capabilities withwhich you are already familiar. Our goal in buildingthe new feature data model has been to move asmuch specialized application logic (for example,maintaining connectivity or relational integritybetween objects) as possible into the scope of thedata model itself. This allows more of the GISapplication to be defined using rules in the data

model, rather than custom application logic writtenfor each application. For other aspects of the datamodel, which may already be familiar to the reader,the specific jargon and concepts used in ArcInfo 8(for topics like image data, as an example) areclearly introduced and defined.

This book also connects the specialized world ofgeographic information systems and the broaderworld of object-relational databases. ArcInfo nowsupports the direct use of standard relationaldatabase technology as an integral part of the GIS.This introduces some new concepts to the GIScommunity. Topics such as transaction models forsimultaneous editing of a shared, seamless databaseare described in detail. For the GIS specialist, thisprovides a good introduction to standard databaseconcepts. For the database specialist, this bookserves as a good answer to the question “what is sospecial about spatial?”

Working with geographic information systems is funfor me because it serves to integrate concepts andideas from a variety of different disciplines—geometry and networks from applied mathematics,sampling and measurement theory from remotesensing and physics, information modeling andmultiuser database issues from informationtechnology. In working with GIS, we get to integrateall of this in a single, useful framework for buildingreal systems. This book presents that synthesis,based on our work with ArcInfo 8. I hope you findthis book useful and stimulating as a basis for yourown work in geographic information systems.

Scott MorehouseDirector of Software DevelopmentEnvironmental Systems Research Institute, Inc.Redlands, California

Acknowledgments

This book, Modeling Our World, is the distillation ofmany people’s inspirations, ideas, and labors.

Many deserve recognition—the ArcInfo usercommunity, which always amazes us with creativeapplications of GIS; the ArcInfo 8 development team,which has produced a true masterpiece of software;and the teams throughout ESRI, which collaboratedto take GIS technology to new levels.

Because of the constraints of space, only a few canbe directly acknowledged. These are some of thecontributors to this software release and book.

The structural design of ArcInfo 8 was led by someof the brightest thinkers in the industry. Sud Menondirected the architectural design of the geodatabaseand he is responsible for many of the insightsexpressed in this book. Jeff Jackson led theimplementation of software component technologythat has revolutionized ArcInfo. Erik Hoel applied hisexpertise to the development of the network featuresand the framework for vertical applications. Thedevelopment of the ArcMap™ and ArcCatalog™applications was led by Barry Michaels, Scott Simon,and Keith Ludwig. The accessibility and consistencyof the software user interface was guided by RupertEssinger. This complex endeavor was orchestratedby Matt McGrath.

Many product specialists and programmers at ESRIprovided material for this book and reviewedchapters. These include Andy MacDonald, CharlieFrye, Mike Minami, Aleta Vienneau, Jim TenBrink,Wolfgang Bitterlich, Tom Brown, Dale Honeycutt,Steve Kopp, Brett Borup, Peter Petri, ClaytonCrawford, and Andrew Perencsik. The contributionsof Andy, Dale, and Steve to chapters 5, 8, and 9respectively are particularly noteworthy.

The attractive city maps throughout this book werekindly provided by Gar Clarke, GIS manager at theCity of Santa Fe, New Mexico. The image of Mars atthe front of chapter 9 is courtesy of Malin SpaceScience Systems and JPL/NASA.

The maps on the chapter title pages are drawn fromthe work of many cartographers from history. Theirmaps remind us that, although we have reached alevel of sophistication in drawing maps withcomputers, we have yet to equal their artistry.

Several people were actively engaged in theproduction of this book. Jennifer Wrightsellrigorously edited the chapters and designed thelayout, along with Andy Mitchell and Youngiee Auh.Amaree Israngkura designed the cover. MichaelHyatt did the copyedit. Robin Floyd and ChristianHarder managed and guided the publication of thisbook.

Scott Morehouse wrote the preface and is ESRI’svisionary on advancing the theory and practice ofGIS. Clint Brown prodded and inspired us to createthe best product we had within ourselves. CurtWilkinson and David Maguire worked hard to ensurethat ArcInfo 8 meets the goals and requirements ofusers. Jack Dangermond created this very specialand unique institute where we can believe that wemake a difference in this world and act on that idea.

Finally, my wife Elizabeth deserves special thanks forher countless hours of support. Her commitment andencouragement made the effort to produce this bookpossible.

Contents

PREFACE ........................................................................................................................... vii

ACKNOWLEDGMENTS ............................................................................................. ix

CHAPTER 1: OBJECT MODELING AND GEODATABASES ........................ 1

Modeling objects with GIS ...................................................................................... 2

The progress of geographic data models .............................................................. 4

The geodatabase, store of geographic data ........................................................... 8

Features in an object-oriented data model ...........................................................10

Serving geographic data .........................................................................................12

Accessing geographic data .....................................................................................14

Building data models ..............................................................................................16

Guide to reading UML object diagrams ................................................................19

Technology trends ..................................................................................................21

CHAPTER 2: HOW MAPS INFORM...................................................................... 23

The utility of maps ..................................................................................................24

How maps present information .............................................................................25

The parts of a map .................................................................................................27

Presenting geography with layers ......................................................................... 28

Drawing features with symbols ............................................................................. 30

Drawing feature layers ............................................................................................32

Classifying attribute values .....................................................................................36

Displaying thematic, spectral, and picture data ....................................................38

Visualizing surfaces with TIN layers .....................................................................41

iv • Modeling Our World

CHAPTER 3: GIS DATA REPRESENTATIONS ................................................... 45

The fundamentals of a GIS ....................................................................................46

The diverse applications of GIS ............................................................................48

Three representations of the world .......................................................................51

Modeling surfaces .................................................................................................. 52

Modeling imaged or sampled data ........................................................................54

Modeling discrete features .....................................................................................56

Comparing spatial data representations ................................................................58

CHAPTER 4: THE STRUCTURE OF GEOGRAPHIC DATA ........................ 61

The catalog and connections to data ....................................................................62

The geodatabase, datasets, and feature classes ...................................................64

ArcInfo workspaces and coverages ..................................................................... 66

Shapefiles and CAD files ........................................................................................68

Maps and layers ......................................................................................................70

Comparing the structure of vector datasets ..........................................................72

Comparing feature geometry in vector datasets ...................................................73

CHAPTER 5: SMART FEATURES ............................................................................. 75

The qualities of features .........................................................................................76

Steps to making features smart ............................................................................. 78

Designing the geodatabase ....................................................................................80

Storing data in tables ..............................................................................................82

The shape and extent of features ......................................................................... 84

Attributes: qualities of an object ............................................................................86

Adding simple behavior with subtypes .................................................................88

Validating attributes .................................................................................................90

Relationships among objects .................................................................................92

Extending object classes ........................................................................................96

The geodatabase object model ............................................................................. 98

CHAPTER 6: THE SHAPE OF FEATURES ........................................................ 101

Geometry and features ......................................................................................... 102

Constructing geometry .......................................................................................... 105

Testing spatial relationships ................................................................................. 110

Applying topological operators ............................................................................ 112

Geometry object model ........................................................................................ 114

Contents • v

CHAPTER 7: MANAGING WORK FLOW WITH VERSIONS ................. 115

Using versions .......................................................................................................116

Long transactions and the geodatabase ..............................................................118

The fundamentals of versions ............................................................................. 120

Editing versioned geodatabases ........................................................................... 122

Types of work flows .............................................................................................124

CHAPTER 8: LINEAR MODELING WITH NETWORKS ........................... 127

Modeling infrastructure ........................................................................................128

The network model ..............................................................................................130

How features connect ..........................................................................................132

Network features ...................................................................................................134

Network flow .........................................................................................................139

Analysis on a network ..........................................................................................142

Network object model ..........................................................................................145

CHAPTER 9: CELL-BASED MODELING WITH RASTERS ........................ 147

Representing geography with rasters ...................................................................148

Using raster data ....................................................................................................150

Raster data model .................................................................................................152

Raster display and analysis ..................................................................................154

The spatial context of rasters ............................................................................... 156

Raster formats ........................................................................................................158

Raster object model ..............................................................................................160

CHAPTER 10: SURFACE MODELING WITH TINS ...................................... 161

Representing surfaces ...........................................................................................162

Structure of a TIN .................................................................................................164

Modeling surface features ....................................................................................166

vi • Modeling Our World

CHAPTER 11: FINDING LOCATIONS............................................................. 169

Using locations ...................................................................................................... 170

Converting locations to map features .................................................................. 172

Converting x,y locations ....................................................................................... 173

Converting addresses ............................................................................................ 174

Converting place names ....................................................................................... 177

Converting postal zones ....................................................................................... 178

Converting route locations ................................................................................... 179

CHAPTER 12: GEODATABASE DESIGN GUIDE .......................................... 181

Purpose and goals of design ............................................................................... 182

Overview of design steps ..................................................................................... 184

Step 1: Model the user’s view .............................................................................. 186

Step 2: Define entities and relationships ............................................................. 188

Step 3: Identify representation of entities ........................................................... 190

Step 4: Match to geodatabase data model .......................................................... 192

Step 5: Organize into geographic data sets ......................................................... 194

INDEX ............................................................................................................................. 197

1

1A geographic data model is a representation ofthe real world that can be used in a GIS toproduce maps, perform interactive queries, andexecute analysis.

Contemporary developments in database andsoftware technology are enabling a newgeneration of geographic data models. Theseare the topics in this chapter:

• Modeling objects with GIS

• The progress of geographic data models

• The geodatabase, store of geographic data

• Features in an object-oriented data model

• Serving and accessing geographic data

• Building data models

• Guide to reading UML object diagrams

• Technology trends

Objectmodeling andgeodatabases

Northern Polar Region. Gerhard Mercator, 1595.

2 • Modeling Our World

The purpose of a geographic information system(GIS) is to provide a spatial framework to supportdecisions for the intelligent use of earth’s resourcesand to manage the man-made environment.

Most often, a GIS presents information in the formof maps and symbols. Looking at a map gives youthe knowledge of where things are, what they are,how they can be reached by means of roads orother transport, and what things are adjacent andnearby. A GIS can also disseminate informationthrough an interactive session with maps on apersonal computer. This interaction revealsinformation that is not apparent on a printed map.

For example, you can query all known attributes ofa feature, create a list of all things connected fromone point on a network to another, and performsimulations to gauge qualities such as water flow,travel time, or dispersion of pollutants.

The way you choose to display and analyzeinformation depends upon how you modelgeographic objects from the world.

MANY WAYS TO MODEL A SYSTEM

Our interaction with objects in the world is diverse,and you can model them in many ways.

Consider one example, rivers. Rivers are naturalfeatures, are used for transportation, delimit politicalor administrative areas, and are an important featurein the shape of a surface. Here are a few of themany ways you can think about modeling riversin a GIS:

• As a set of lines that form a network. Eachsection of line has flow direction, volume, andother attributes of a river. You can apply a linearnetwork model to analyze hydrographic flow orship traffic.

• As a border between two areas. A river candelimit political areas such as provinces orcounties, or can be a barrier for natural regionssuch as wildlife habitats.

• As an areal feature with an accuraterepresentation of its banks, braids, andnavigable channels on the river.

• As a sinuous line forming a trough in a surfacemodel. From the river’s path through a surface,you can calculate its profile and rate of descent,the watershed it drains, and its flooding potentialfor a prescribed rainfall.

MAP USE GUIDES THE DATA MODEL

It is clear that even a common type of geographicfeature such as a river can be represented in a GISin a variety of ways. No model is intrinsicallysuperior; the type of map you want to create andthe context of the problems to be solved will guidewhich model is best.

MODELING OBJECTS WITH GIS

Chapter 1 • Object modeling and geodatabases • 3

�

�

The geodatabase stores locations �such as addresses, x,y locations, �postal codes, place names, and route �locations. Locators contain�information to create features.

A network is a set of features that �participate in a linear system such�as a utility network, stream network,�or road network. Networks are well�suited for tracing analysis.

Raster technology is an efficient �means of capturing large amounts �of imaged data. Images provide an�informative background display �below feature layers on a map.

Features are discrete objects on a �map. Small objects are represented �as points, long objects as lines, and �broad objects as polygons.

locationimage

surface

network

227 East Palace Avenue

features

The earth’s surface can be kept in�a geodatabase in several forms: as�a triangulated irregular network (TIN),�as elevation values on cells in�a raster, or as contour lines.

Representations of geography

�


THE PROGRESS OF GEOGRAPHIC DATA MODELS

A geographic data model is an abstraction of thereal world that employs a set of data objects thatsupport map display, query, editing, and analysis.

ArcInfo 8 introduces a new object-oriented datamodel—the geodatabase data model—that iscapable of representing natural behaviors andrelationships of features. To understand the impactof this new model, it is instructive to review threegenerations of geographic data models.

THE CAD DATA MODEL

The very first computerized mapping systems drewvector maps with lines displayed on cathode raytubes and raster maps using overprinted characterson line printers. From this genesis, the 1960s and1970s saw the refinement of graphics hardware andmapping software that could render maps withreasonable cartographic fidelity.

In this era, maps were usually created with general-purpose CAD (computer-aided design) software.The CAD data model stored geographic data inbinary file formats with representations for points,lines, and areas. Scant information about attributeswas kept in these files; map layers and annotationlabels were the primary representation of attributes.

THE COVERAGE DATA MODEL

In 1981, Environmental Systems Research Institute,Inc. (ESRI), introduced its first commercial GISsoftware, ArcInfo, which implemented a second-generation geographic data model, the coveragedata model (also known as the georelational datamodel). This model has two key facets:

• Spatial data is combined with attribute data. Thespatial data is stored in indexed binary files,which are optimized for display and access. Theattribute data is stored in tables with a number ofrows equal to the number of features in thebinary tables and joined by a common identifier.

• Topological relationships between vector featurescan be stored. This means that the spatial datarecord for a line contains information aboutwhich nodes delimit that line, and by inference,which lines are connected; it also containsinformation about which polygons are on itsright and left sides.

The major advance of the coverage data model wasthe user’s ability to customize feature tables; notonly could fields be added, but database relatescould be set up to external database tables.

ArcPolygon

Labelpoint

Polygon Attribute Table

Arc Attribute Table

Point Attribute Table

Coverage Attributes in relational tables

Spatial data in relational tables

Because of the performance limitations of computerhardware and database software of the time, it wasnot practical to store spatial data directly in arelational database. Rather, the coverage data modelcombined spatial data in indexed binary files withattribute data in tables.

Despite this compromise of partitioning spatial andattribute data, the coverage data model has becomethe dominant data model in GIS. This has been forgood reason—the coverage data model made high-performance GIS possible, and stored topologyfacilitated improved geographic analysis and moreaccurate data entry.

Limitations of the coverage data model

However, the coverage data model has an importantshortcoming—features are aggregated intohomogeneous collections of points, lines, andpolygons with generic behavior. The behavior of aline representing a road is identical to the behaviorof a line representing a stream.

The generic behavior supported by the coveragedata model enforces the topological integrity of adataset. For example, if you add a line across apolygon, it is automatically split into two polygons.

But it is desirable to also support the specialbehaviors of streams, roads, and other real-worldobjects. An example is that streams flow downhilland when two streams merge into one, the flow ofthe merged stream is the addition of the twoupstream flows. Another example is that when tworoads cross, a traffic intersection should be at theirjunction unless there is an overpass or underpass.


Customizing features in coverages

With the coverage data model, ArcInfo applicationdevelopers had some notable success in adding thistype of behavior to features through macro codewritten in the ARC Macro Language (AML™). Manysuccessful, large-scale, industry-specific applicationswere built.

However, as applications became more complex, itbecame apparent that a better way to associatebehavior with features was needed. The problemwas that the developer had the task of keeping theapplication code in synchronicity with featureclasses—no easy task. The time had come for anew geographic data model with an infrastructureto tightly couple behavior with features.

THE GEODATABASE DATA MODEL

ArcInfo 8 introduces a new object-oriented datamodel called the geodatabase data model. Thedefining purpose of this new data model is to letyou make the features in your GIS datasets smarterby endowing them with natural behaviors, and toallow any sort of relationship to be defined amongfeatures.

The geodatabase data model brings a physical datamodel closer to its logical data model. The dataobjects in a geodatabase are mostly the sameobjects you would define in a logical data model,such as owners, buildings, parcels, and roads.

Further, the geodatabase data model lets youimplement the majority of custom behaviors withoutwriting any code. Most behaviors are implementedthrough domains, validation rules, and otherfunctions of the framework provided in ArcInfo.Writing software code is only necessary for themore specialized behaviors of features.

SCENARIOS OF OBJECT INTERACTIONS

To get a sense of why an object-oriented data modelis important, review the following scenarios thatillustrate common tasks you might perform withfeatures. From these scenarios, you can sift out thebenefits of an object-oriented data model and thenreview some specific characteristics of thegeodatabase data model.

Adding and editing features

When you add geographic features to your GISdatabase, you want to ensure that features areplaced correctly according to rules such as these:

• That the values you assign to an attribute fallwithin a prescribed set of permissible values. Aparcel of land may only have certain land usessuch as residential, agricultural, or industrial.

residential�agricultural�commercial�industrial

table

row

column

• That a feature can be placed adjacent orconnected to another feature only if certainconstraints are met. Placing a liquor store near aschool is not permitted by law. A city roadcannot be connected to a highway without atransition segment such as an on-ramp.

highway

transition

road

• That collections of certain features conform totheir natural spatial arrangement. A stream systemshould always flow downhill. Flow down from ajunction is the sum of flows upstream.

• That the geometry of a feature follows its logicalplacement. The lines and curves that make up aroad should be tangent. Building corners mostoften form right angles.


Relationships among features

All objects in the world are entangled inrelationships with other objects. From theperspective of a GIS, these relationships can beconsidered to fall within three general categories:topological, spatial, and general.

These are some examples of each of these types ofrelationships:

• When you edit features in an electric utilitysystem, you want to be sure that the ends ofprimary and secondary lines connect exactly andthat you are able to perform tracing analysis onthat electric network. A set of topologicalrelationships is defined for you when you loador edit features within a connected system.

• When you work with a map with buildings,blocks, and school districts, you might want todetermine which block contains a particularbuilding, the set of all buildings within a schooldistrict, and which blocks contain no buildings.A fundamental function of a GIS is to determinewhether a feature is inside, touching, outside, oroverlapping another feature. Spatial relationshipsare inferred from the geometry of features.

• Some objects have relationships that are notpresent on a map. A parcel has a relationship toan owner, but the owner is not a feature on amap. A general relationship connects the parceland the owner. Some features on a map haverelationships, but their spatial relationship isambiguous. A utility meter is in the generalvicinity of an electric transformer, but it is nottouching the transformer. The meter and the

transformer might not be reliably related by theirspatial proximity in crowded areas, so a generalrelationship ties the two features together.

transformermeter

parcel

owner

Cartographic display

Most of the time, you will draw features on a mapwith predefined symbols, but sometimes you willwant more control over how your features aredrawn. These are some specialized drawingbehaviors:

5280

5280

• When you display a contour line, you want itselevation annotated along a flat section of thecontour, at an average interval such as 4 inches,and not obscuring other features.

• When you draw roads on a detailed map, youwould like the road drawn as parallel lines withclean intersections wherever there is a roadintersection.

Circuit C

Circuit ACircuit B

Pole

• When multiple electrical wires are physicallymounted on the same set of utility poles, youwould like to depict them as spread in a set ofparallel lines with a standard offset in map units.


Interactive analysis

Dynamic map displays invite the user to touchfeatures, find properties and relationships, andlaunch analyses. These are examples of some tasksyou may want to perform upon selected features:

• Touch a feature on a map display and invoke aform to query and update its properties.

• Select a part of an electric network where linemaintenance is planned, find all affecteddownstream customers, and make a mailing listto notify them.

BENEFITS OF THE GEODATABASE DATA MODEL

The common thread throughout these scenarios isthat it is very useful to apply object-oriented datamodeling to features. Object-oriented data modelinglets you characterize features more naturally byletting you define your own types of objects, bydefining topological, spatial, and generalrelationships, and by capturing how these objectsinteract with other objects. Some of the benefits ofthe geodatabase data model are:

• A uniform repository of geographic data. All ofyour geographic data can be stored and centrallymanaged in one database.

• Data entry and editing is more accurate. Fewermistakes are made because most of them can beprevented by intelligent validation behavior. Formany users, this alone is a compelling reason toadopt the geodatabase data model.

• Users work with more intuitive data objects.Properly designed, a geodatabase contains dataobjects that correspond to the user’s model of

data. Instead of generic points, lines, and areas,the users work with objects of interest, such astransformers, roads, and lakes.

• Features have a richer context. With topologicalassociations, spatial representation, and generalrelationships, you not only define a feature’squalities, but its context with other features. Thislets you specify what happens to features whena related feature is moved, changed, or deleted.This context also lets you locate and inspect afeature that is related to another.

• Better maps can be made. You have more controlover how features are drawn and you can addintelligent drawing behavior. You can applysophisticated drawing methods directly in theArcInfo mapping application, ArcMap. Highlyspecialized drawing methods can be executed bywriting software code.

• Features on a map display are dynamic. Whenyou work with features in ArcInfo, they canrespond to changes in neighboring features. Youcan also associate custom queries or analytictools with features.

• Shapes of features are better defined. Thegeodatabase data model lets you define theshapes of features using straight lines, circularcurves, elliptical curves, and Bézier splines.

• Sets of features are continuous. By their design,geodatabases can accommodate very large setsof features without tiles or other spatial partitions.

• Many users can edit geographic datasimultaneously. The geodatabase data modelpermits work flows where many people can editfeatures in a local area, and then reconcile anyconflicts that emerge.

To be sure, you can realize some of these benefitswithout an object-oriented data model, but youwould be at a disadvantage—you would need towrite external code loosely coupled to features andprone to complexity and error. A principaladvantage of the geodatabase data model is that itincludes a framework to make it as easy as possibleto create intelligent features that mimic theinteractions and behaviors of real-world objects.


A geodatabase can contain four representations ofgeographic data:

• Vector data for representing features

• Raster data for representing images, griddedthematic data, and surfaces

• Triangulated irregular networks (TINs) forrepresenting surfaces

• Addresses and locators for finding a geographicposition

A geodatabase stores all of these representations ofgeographic data in a commercial relationaldatabase. This means that geographic data can beadministered centrally by information technologyprofessionals and ArcInfo can take advantage ofdevelopments in database technology.

REPRESENTING FEATURES WITH VECTORS

Many of the features in the world have well-definedshapes. Vector data represents the shapes offeatures precisely and compactly as an ordered setof coordinates with associated attributes. Thisrepresentation supports geometric operations suchas calculating length and area, identifying overlapsand intersections, and finding other features that areadjacent or nearby.

Vector data can be classified by dimension:

• Points are zero-dimensional shapes representinggeographic features too small to be depicted aslines or areas. Points are stored as a single x,ycoordinate with attributes.

• Lines are one-dimensional shapes that representgeographic features too narrow to depict asareas. Lines are stored as a series of ordered x,ycoordinates with attributes. The segments of aline can be straight, circular, elliptical, or splined.

• Polygons are two-dimensional shapes thatrepresent broad geographic features stored as aseries of segments that enclose an area. Thesesegments form a set of closed areas.

Another type of vector data is annotation. These aredescriptive labels that are associated with featuresand display names and attributes.

THE GEODATABASE,STORE OF GEOGRAPHIC DATA

Vector data in a geodatabase has a structure thatdirects the storage of features by their dimensionand relationships. A feature dataset is the containerof spatial entities (features) and nonspatial entities(objects) and the relationships between them.Topological associations are represented withgeometric networks and planar topologies.

A geodatabase also stores validation rules anddomains to ensure that when features are created orupdated, their attributes remain valid in the contextof related features and objects.

REPRESENTING GRIDDED DATA WITH RASTERS

Much of the data collected in a geodatabase is ingrid form. This is because cameras and imagingsystems record data as pixel values in a two-dimensional grid, or raster.

A cell is a pixel element of a raster and its valuescan depict a variety of data. A cell can store thereflectance of light for part of the spectrum, a colorvalue for a photograph, a thematic attribute such asvegetative type, a surface value, or elevation.

REPRESENTING SURFACES WITH TINS

A triangulated irregular network (TIN) is a model ofa surface. A geodatabase stores TINs as anintegrated set of nodes with elevations and triangleswith edges. An elevation (or z value) can beinterpolated for any point within the geographicextent of a TIN.

TINs enable surface analysis such as watershedstudies, visibility of a surface from an observationpoint, and delineation of surface features such asridges, streams, and peaks. TINs can also depict thephysical relief of terrain.

Note: At the initial release of ArcInfo 8, a geodatabasedoes not yet store TINs or rasters. For the interim, TINscan be stored in coverage workspaces and rasters infolders or workspaces.

FINDING ADDRESSES WITH LOCATORS

Perhaps the most common geographic task isfinding an address. A geodatabase can storeaddresses and other locations. Geodatabases alsostore locators containing information that allowsyou to create features for locations.


Geodatabase

Feature datasets

Spatial reference

Geometric networks

Planar topologies

Domains

Raster datasets

Rasters

Validation rules

Locators

Addressesx,y locationsZIP CodesPlace namesRoute locations

77 Sunset

rastersurface

vector

location

Raster datasets can represent an imaged map, a surface,an environmental attribute sampled on a grid, orphotographs of objects referenced to features. Someraster data is collected in bands that commonly representdifferent spectral ranges of camera filters.

TIN datasets are triangulations of sets of irregularly locatedpoints with z-values (elevations) sampled from a surface.TINs are most often used to model the earth’s surface, butare also used to study the distribution of a continuousenvironmental factor such as chemical concentration.

Corporate and agency databases have many records withaddresses and other locations. Locators containinformation that allows you to create features for locationsso you can display them on a map.

topology

data integrity

entitiesrelationships

Feature classes, subtypes

Object classes, subtypes

Relationship classes

A feature dataset contains objects and features and therelationships among them. An object is a nonspatial entityand a feature is a spatial entity. A relationship links twoentities.

Objects of the same kind are stored in an object class.Features of the same kind and with the same type ofgeometric shape are stored in a feature class.

A relationship class stores relationships between entities intwo object or feature classes.

A

Geometric networks model linear systems such as utilitynetworks and transportation networks. They support a richset of network-tracing and -solving functions.

Domains are sets of valid attribute values for objectattributes. They can be textual or numeric.

Validation rules enforce data integrity through relationshiprules and connectivity rules.

can be

inside or

outside of

feature

datasets

spatial referenceAll feature classes in a feature dataset share a commoncoordinate system. Because the feature dataset is thecontainer of topological associations, it is important toguarantee a common spatial reference.

Planar topologies model systems of line and area featuresas a continuous coverage of an area. Planar topologiesallow features to share common boundaries, such ascounties sharing an outer boundary with a state.

TIN datasets

nodesedges

faces

Inside a geodatabase


FEATURES IN AN OBJECT-ORIENTED DATA MODEL

ArcInfo 8 is distinguished from antecedent releasesas it applies object-oriented methodology togeographic data modeling. A developer interactswith data objects through a framework of object-oriented software classes called the geodatabasedata access objects.

There are three key hallmarks of object orientation:polymorphism, encapsulation, and inheritance.

• Polymorphism means that the behaviors (ormethods) of an object class can adapt tovariations of objects. For example, the corebehavior of features, such as draw, add, anddelete operations, is the same whether thefeatures reside in a geodatabase, coverage, orshapefile.

• Encapsulation means that an object is accessedonly through a well-defined set of softwaremethods, organized into software interfaces. Thegeodatabase data access objects mask theinternal details of data objects and provide astandard programming interface.

• Inheritance means that an object class can bedefined to include the behavior of anotherobject class and have additional behaviors. Youcan create custom feature types in ArcInfo andinherit the behavior of standard features. Forexample, a transformer object can be extended(or subtyped) from a standard ArcInfo featuretype such as a simple junction feature.

UNIFIED DATA MODEL

The geodatabase data access objects comprisesoftware technology that provides uniform access togeographic data from several data sources such asgeodatabases, coverages, and shapefiles.

ArcInfo developers interact with geographic datathrough a set of data items, such as datasets, tables,feature classes, rows, objects, and features. Theseitems comprise a common and consistent view ofgeographic data.

Because of this unified data model, the ArcInfouser can work with geodatabases, coverages, andshapefiles in the same way. The unified data modelsimplifies how users work with data by emphasizingthe common characteristics of data.

EXTENSIBLE FEATURES

An important aspect of a geodatabase is that youcan optionally create custom features such astransformers and roads, instead of points and lines.

To the ArcInfo user, this means that a transformeror road has all of the standard display, query, andedit behavior of standard point features and linefeatures, but with additional behaviors. You canspecify that a transformer must be drawn touchinga power pole and perpendicular to the electric linethrough the pole. Or, when a road is edited, all ofits segments must be tangent.

A data modeler can use standard feature types toimplement a rich data model. For advancedapplications, a developer can extend the standardfeature types and create custom features using theobject-oriented technique of type inheritance.

Any custom feature that you create enjoys the sameperformance and functionality as the standardfeature types provided by ArcInfo. This offerslimitless opportunities for sophisticated applicationdevelopment.

FEATURES AND OBJECT ORIENTATION

Features in a geodatabase are implemented as a setof relational tables. Some of these tables representcollections of features. Other tables representrelationships between features, validation rules, andattribute domains.

ArcInfo manages the structure and integrity of thesetables and presents an object-oriented geographicdata model through the geographic data accessobjects.

All users and most developers will not know or careabout the details of the internal structure of ageodatabase. The ArcCatalog application is youruser interface to establish, modify, and refine thestructure of your geodatabase.

The object view of data lets you focus your effortson building a geographic data model and hidesmost of the physical database structure of thegeodatabase.


datacomponents

datasources

ArcInfoapplications

geodatabase data access objects

geodatabase coverage shapefile

Junction-Feature

Simple-Junction-Feature

Complex-Junction-Feature

Transformer Switch

customfeature

types

standardfeature

types

Data can be viewed in three ways.��

The relational table view of data�exposes the internal details of the�

physical storage as database tables.��

The simple feature view presents data�in the form of features without the�

structure of topology and relationships.��

The object view of data encapsulates� the internal details and presents a�

higher level of structure that is closer� to the user’s conceptual model of data.�

��

The geodatabase data access objects include�a number of software components that represent�

the types of features that are ready for use.�Shown here are some of the network feature�

types. These have intrinsic behaviors that�guarantee the topological integrity of features� in a geometric network. Most data modelers�use standard feature types without extending�

them through custom programming.

ArcInfo is versatile at�displaying and analyzing�

geographic features.�ArcInfo works with a�

number of data sources,�including geodatabases,�

coverages, and shapefiles.

Features in a geodatabase

The geodatabase data�access objects comprise a�

programming interface�that largely hides any�

differences among feature�types from geodatabases,�

coverages, and shapefiles.

These are some custom features that� have been extended from the standard�

feature types. They implement�specialized behaviors for custom�

applications developed by data modelers� and programmers.

unifieddata model

extensiblefeatures

data access

polymorphism

inheritance

encapsulationFeature-Dataset Table

Dataset

Relationship-ClassObjectClass

Feature-Class

object view of data

relational table

�geometry

column

rules, domains

relationships

attributecolumns

relational tableview of data

simple feature view of data

points

lines

polygons

geometric shapes with attributes

SwitchGear-Cabinet


SERVING GEOGRAPHIC DATA

ArcInfo accesses geographic data served throughArcSDE™, the Arc Spatial Database Engine. ArcSDEis the software technology that enables you tocreate geodatabases that range from small to verylarge sets of geographic data, and provides an openinterface to the relational database of your choice.

HOW A GEODATABASE EXTENDS A DATABASE

These are some of the facets of a geodatabase thatenhance relational database technology:

• A geodatabase can represent geographic data infour manifestations: discrete objects as vectorfeatures, continuous phenomena as rasters,surfaces as TINs, and references to places aslocators and addresses.

• A geodatabase stores shapes of features andArcInfo provides functions for performing spatialoperations such as finding objects that arenearby, touching, or intersecting. A geodatabasehas a framework for defining and managing thegeographic coordinate system for a set of data.

• A geodatabase can model topologicallyintegrated sets of features such as transportationor utility networks and subdivisions of landbased on natural resources or land ownership.

• A geodatabase can define general and arbitraryrelationships between objects and features.

• A geodatabase can enforce the integrity ofattributes through domains and validation rules.

• A geodatabase can bind the natural behavior offeatures to the tables that store features.

• A geodatabase can present multiple versions sothat many users can edit the same data.

PERSONAL AND MULTIUSER GEODATABASES

Geodatabases comes in two variants—personal andmultiuser.

Personal geodatabase support is built into ArcInfoand is suitable for project-oriented GIS. A personalgeodatabase is implemented as a Microsoft® Accessdatabase. When you install ArcInfo, Microsoft Jet isalso installed; this provides the services for ArcInfoto create and update Access databases. You do notneed to separately install Microsoft Access.

For large enterprises, you can deploy multiusergeodatabases with ArcSDE—the multiuser dataaccess extension to ArcInfo. ArcSDE is installed ona data server that administers your organization’srelational database. Through a TCP/IP network,ArcSDE serves geodatabases to the ArcInfoapplications running on personal computers.ArcSDE can be run on Windows NT® or UNIX®.

ArcSDE allows remote access to geographic dataand allows many users to view and edit the samegeographic data. ArcSDE is centrally tuned andmanaged by your database administrator.

AN OPEN AND SCALABLE DATA SERVER

ArcInfo allows you to configure and deploy smallto very large geodatabases. If you are working withmoderately sized datasets, you can deploy personalgeodatabases in ArcCatalog. This configurationyields good performance for datasets up toapproximately 250,000 objects and supports oneeditor and several simultaneous viewers.

For more demanding datasets and to support manyconcurrent editors, you can deploy ArcSDE on therelational database best suited to yourorganization’s needs.

These are some reasons to add ArcSDE to yourArcInfo installation:

• You have limitless flexibility in scaling databases.

• You can deploy the relational database of yourchoice.

• You can serve geographic data from UNIX orWindows NT.

• You can serve data to other applications such asMapObjects®, ArcIMS™ (Arc Internet Map Server),ArcView® GIS, and CAD client applications.

• You can centrally store and administergeodatabases.

• You can build Open GIS Consortium (OGC)-compliant applications.

• You can build Structured Query Language (SQL)applications to access the tables and rows in ageodatabase.


Geodatabase

Feature dataset

Feature class

Feature class

Relationship class

Geometric network

Feature class

Geodatabase

Feature dataset

Feature class

Feature class

Feature class

Object class

Relationship class

Geodatabase

Object class

Relationship class

Feature class

Feature class

Feature class

Object class

geographically enhanceddatabases

Geodatabase

Raster datasetRaster

Feature class


A personal geodatabase isdirected toward personal orsmall work-group use. It canhandle small to moderately

sized datasets.

A geodatabase served through ArcSDE can manage very large sets of geographicdata and serve large numbers of viewers and editors. Geographic data is

accessed from a data server on a network. This GIS data is centrally administeredin large databases and integrates well with other corporate data. These databases

require a system administrator for permissions, tuning, and optimization.

Personal geodatabases areimplemented on the Microsoft

Jet engine, which stores data asMicrosoft Access databases.

ArcSDE operates on any leading relational database. The ArcInfo developer caninteract with a geodatabase through the geodatabase data access objects. A

developer can access an ArcSDE geodatabase outside of ArcInfo through a C API(application programmer interface) or an SQL API.

To model work-flow processes, a geodatabase served through ArcSDE supportslong transactions and version management. A versioned geodatabase allowsmany editors to work concurrently and includes a framework for resolving edit

conflicts.

A personal geodatabase has all thefunctionality of a geodatabase

served through ArcSDE, except forversioning.

Geodatabases on any supported relational database operate identically.

relational databases

Personalgeodatabase

Personalgeodatabasesupport is builtinto ArcInfo andprovides accessto local data.

ArcSDE

ArcSDE is a technology that uses thenative data types and operators in arelational or object-relational databaseand extends them to provide the completefunctionality of a geodatabase.

project GIS enterprise GIS

Geodatabase

LocatorAddresses

TIN dataset

Object class

A geodatabase is aninstance of a relational or

object-relational databasethat has been enhanced

by adding geographic datastorage, referential

integrity constraints, mapdisplay, feature-editing,and analysis functions.

ArcSDE is the multiuser extension to ArcInfo.

MicrosoftAccess

Oracle 8SQL

ServerInformix DB2 Sybase

Open data framework


ACCESSING GEOGRAPHIC DATA

A developer can access data in a geodatabase atthree basic levels:

• Through the geodatabase data access objects, asubset of ArcObjects™, the software componentson which ArcMap and ArcCatalog are built.

• As simple nontopological features through theArcSDE application programmer interface thatcomplies with the OGC simple featurespecification.

• As raw rows, columns, and tables through thenative SQL interface of the relational database.

ACCESSING DATA THROUGH ARCOBJECTS

The richest level of accessing data is through thegeodatabase data access objects. At this level, the fullstructure of a geodatabase is revealed: topology,relationships, integrity rules, and behavior, as well asraster, surface, and location representations.

You can programmatically access data throughArcObjects using Microsoft Visual Basic® forApplications (VBA) or with Visual C++® or otherCOM-compliant development environment.

The following is a simplified Unified ModelingLanguage (UML) diagram of a portion of thegeodatabase data access objects. This is discussedin chapter 4, “The structure of geographic data.”

Feature-Dataset

Table

Geometric-Network

GeoDataset

Rule

1..*

1..* 1..*

Attributed-Relationship-

Class

Relationship-Class

ObjectClass

Feature-Class

Graph

WorkspaceDomain

Dataset

Raster-Dataset

TinDataset

0..1

ACCESSING DATA AS SIMPLE FEATURES

For many spatial applications, it is sufficient anddesirable to access geographic data in the form ofsimple nontopological features. This approach isespecially suitable for building integratedapplications for which geographic data is a vitalcomponent, but perhaps not the focus. Examplesinclude facilities management and traffic analysis.

ArcSDE presents a simple feature API in C andJava™ that is compliant with the OGC simplefeatures specification.

OGC is an organization of leading spatial datavendors, and its purpose is to develop standardsoftware interfaces for the free exchange of spatialinformation among heterogeneous GISs.

Organizations that have geographic data in variousformats on a network can build applications thatintegrate this data in the form of simple features.

ESRI is a leading contributor to the OGC technicalspecifications and is committed to the openexchange of geographic data.

ACCESSING DATA THROUGH SQL

A GIS is a rich repository of data about naturalfeatures or facilities such as transportation or utilitynetworks. While this data is collected and managedas a geodatabase, external database applicationscan effectively access and share this data fornongeographic use.

Using the native SQL interfaces of your relationaldatabase, you can build applications to mine datafrom your geodatabases and use them for taskssuch as managing inventory, processing workorders, or statistical analysis.

In this view, a geodatabase is a set of tables, rows,and columns. Through the SQL interfaces, you cansee the internal database structure of a geodatabase,which includes metadata tables for objects such asnetworks. This structure is not directly visible inArcInfo and is managed through the user interfaceof ArcCatalog. You can selectively update attributesof rows that represent features, but you should takecare not to corrupt the structure of the geodatabase.


Accessing geodatabases

ArcInfo is a general-purpose GIS application withadvanced editing and map display, spatial analysis, andtopological processing. Through ArcInfo, features inyour geodatabase act with full object awareness asexpressed with domains, validation rules, and customcode. The developer uses the geodatabase dataaccess objects in Visual Basic, Visual C++, or otherCOM-compliant development environments.

relational databases

Personal geodatabase ArcSDE

GIS applications

spatial applications

MicrosoftAccess

Oracle 8SQL

ServerInformix DB2 Sybase

through thegeodatabasedata accessobjects in

ArcInfo andArcSDE

throughArcSDE API

compliantwith

OGC simplefeatures

through SQLinterface inrelationaldatabases

Some applications process spatial queries from a largegeodatabase and serve highly specialized functions.Examples are emergency response and businesslocation. Geodatabases can be accessed as simplefeatures though the ArcSDE simple feature API. Thisincludes both C and Java APIs. The ArcSDE simplefeature API is open and adheres to the OGC simplefeature specification. This allows features in ageodatabase to be used outside of ArcGIS applications.

Database applications sometimes need to extract datafrom a geodatabase, but not to display or spatiallyprocess that data. An example would be to pull or joinutility pole attributes from a geodatabase to a relationaldatabase so that an inventory can be taken. Thedatabase programmer can interact with the tables in ageodatabase through the native SQL interfaces. Thedeveloper should refrain from modifying any geographicshapes or geodatabase system tables.

database applications

Geometry

Multipoint

Segment

CircularArc

CurvePoint

EllipticArc

BézierCurve

Line

Polycurve

Polyline

Polygon

Ring

Path

Spatialapplicationdeveloper

Geodatabase

Object class

Relationship class

Feature class

Feature class


ArcInfodeveloper

Databasedeveloper

Developers can access a geodatabase through thegeodatabase data access objects in ArcInfo, through APIsthat expose simple features, or by the internal tables.Geodatabase


Designing a geodatabase is fundamentally the sameas designing any database. That is because ageodatabase is an instance of a relationaldatabase—one that contains a structure forrepresenting geographic data.

The geodatabase extends, yet simplifies, the designprocess by presenting an object-oriented datastructure that expresses the spatial and topologicalrelationships of geographic features. Part of thisstructure is a special facility for representing sets ofobjects as integrated systems—such as stream androad networks or sets of land parcels. This structureon a set of features is called topology.

The geodatabase data model is the bridge betweenpeople’s cognitive perception of the objectssurrounding them in the world and how thoseobjects are stored in relational databases.

GEODATABASE DESIGN

Traditional relational database design spans twobasic steps—the articulation of a logical data modeland the physical implementation of databasemodels (or schemas).

The logical data model captures the user’s view ofdata and the database model implements the datamodel within the framework of relational databasetechnology.

Designing a logical data model

The key task in building a logical data model is toprecisely define the set of objects of interest and toidentify the relationships between them.

Some examples of objects you might consider arestreets, parcels, owners, and buildings. Someexamples of their relationships are “located at,”“owned by,” and “is part of.”

Once you have an initial logical data model, youcan validate it against the user’s requirements forentering, updating, and accessing data and bytesting it against the organization’s practices andprocedures (or business rules).

It is especially important to involve representativesfrom each prospective user group. A logical datamodel built for a subset of users is guaranteed tohave deficiencies for overlooked users.

Building a logical data model is an iterative processand an art that is acquired through experience.There is no single “correct” model, but there aregood models and bad models. It is difficult todetermine precisely when a model is correct andcomplete, but an indication that you are comingclose is when you can answer “yes” to thesequestions:

• Does the logical data model represent all datawithout duplication?

• Does the logical data model support anorganization’s business rules?

• Does the logical data model accommodatedifferent views of data for distinct groups ofusers?

Representing logical data models

In the past, logical data models were often drawn inwhat are known as entity-relationship diagrams. Anumber of leading object-oriented modelers putforward various design methodologies and diagramnotations.

These methodologies emphasized different aspectssuch as data flow or use-case scenarios, but aproblem with entity-relationship diagrams is thattheir appearance varied with the designmethodology.

More recently, most object-oriented modelers haveadopted the Unified Modeling Language (UML),which is a standard notation for expressing objectmodels and is endorsed by leading software anddatabase companies.

It is important to note that UML is not a designmethodology, but rather a diagrammatic notation.With UML, you can adopt the object-orienteddesign methodology of your choosing and expressthe model in a standard way.

This book uses UML for drawing that ArcInfo objectmodel, called ArcObjects, and for drawing thecustom object models you can create in ageodatabase.

BUILDING DATA MODELS


Implementing a physical database model

A physical database model is built from the logicaldata model. Typically, a specialist in relationaldatabases receives the logical data model from thedata modeler and uses the database administrationtools to define the database schema and create newdatabases ready for data transfer and entry.

The physical database design has some similarity tothe logical data model, but there are differences.Classes of objects may be split or joined whenimplemented in tables. Rules and relationships canbe expressed in several ways.

An important benefit of the geodatabase is that it isa physical implementation of data, but lets youstructure your data in a fashion that is close to thelogical data model.

Elements of the logical and database models

These are the basic elements of the logical datamodel and their corresponding database elements:

Object

Attribute

Class

Row

Column, Field

Table

Database elementsLogical elements

A logical data model is an abstraction of the objectsthat we encounter in a particular application. Thisabstraction is converted into database elements.

An object represents an entity such as a house,lake, or customer. An object is stored as a row.

An object has a set of attributes. Attributescharacterize the qualities of an object, such as itsname, a measure, a classification, or an identifier(or key) to another object. Attributes are stored in adatabase in columns (or fields).

A class is a set of similar objects. Each object in aclass has the same set of attributes. A class is storedin a database as a table. The rows and columns ina table form a two-dimensional matrix.

Handling complex data

Relational databases enjoy their commercialdominance because they implement a simple,elegant, and well-understood theory. This simplicityis at once a strength and a weakness—it isconceptually straightforward to build relationaldatabases, but difficult to model complex data.

Geographic databases contain complex data. Theshapes of line and area features are structured setsof coordinates that cannot be well represented withstandard atomic field types such as integer, real,and string. Further, features are gathered intosystems that have explicit topological relationships,implicit spatial relationships, or generalrelationships.

The relational database is the foundation for ageodatabase. A key purpose of the geodatabase isto handle complex geographic data with a uniformdata model independent of the relational databaseunderneath.

Chapter 12, “Geodatabase design guide,” returns tothese topics in the context of designing andbuilding geodatabases.

logical data modelreality

database implementation

Building

Landparcel

Person building

person

ownership

parcel

A logical data model is constructedto represent the objects of interest

to an application. From the logical datamodel, a database model

is built in a relationaldatabase.


GUIDELINES FOR GEODATABASE DESIGN

The structure of a geodatabase—feature datasets,feature classes, topological groupings, relationships,and other elements—lets you design geographicdatabases that are close to their logical data models.For a data modeler, this is the essential reason forthe introduction of geodatabases into ArcInfo 8.

These are the basic steps in designing a geodatabase:

1 Model the user’s view of data. Perform interviewswith users, understand an organization’sstructure, and analyze the business requirements.

2 Define objects and relationships. Build the logicaldata model with the set of objects, knowing howthey are related to one another.

3 Select geographic representation. Determinewhether vector, raster, surface, or locationrepresentation is best for the data of interest.

4 Match to geodatabase elements. Fit the objects inthe logical data model into the elements of ageodatabase.

5 Organize geodatabase structure. Build thestructure of a geodatabase with consideration ofthematic groupings, topological associations, anddepartment responsibility of data.

This topic is discussed in greater detail inchapter 12, “Geodatabase design guide.”

Building

Landparcel

Person

Geodatabase

Feature dataset

Geometric network

Feature class

12

34

5

Model the user’s view of data.

Define objects and relationships.

Select geographic representation.

Match to geodatabase elements.

Organize geodatabase structure.

Identify organizational functions.Determine data needed to support functions.Organize data into logical groupings.

Identify and describe objects.Specify relationships between objects.Document model in diagram.

Represent discrete features with points, lines, areas.Characterize continuous phenomena with rasters.Model surfaces with TINs or rasters.

Determine geometry type of discrete features.Specify relationships between features.Implement attribute types for objects.

Organize systems of features.Define topological associations.Assign coordinate systems.Define relationships and rules.

Steps to building a geodatabase


GUIDE TO READING UML OBJECT DIAGRAMS

You can approach ArcInfo in two ways: as a userof applications such as ArcMap and ArcCatalog, oras a software developer building customapplications.

Data modelers straddle these two worlds—you usethe applications for most of your work in creatinggeodatabases, but you will sometimes write softwarecode, especially if you are trying to create rich datamodels that support powerful applications.

One aim of this book is to present the importantdata-modeling concepts both as they are applied inthe ArcInfo applications and in the ArcInfosoftware components, called ArcObjects.

A pattern throughout this book is to first present theconcepts for a topic as you experience it through theArcInfo application. Next, that topic is summarizedwith an annotated diagram of the relevant section ofthe ArcInfo object model diagram.

For example, the topic of the structure ofgeodatabases, feature datasets, and feature classesis first discussed from the user’s perspective withinArcCatalog. Next, the programmer’s perspective issummarized in a diagram of part of thegeodatabase data access objects.

These two views have similarities, but also subtledifferences. A user interface sometimes hides detailsabout software components that are important to theprogrammer. One goal of this book is to give youthe insight to bridge the user and developerperspectives.

Reading the class diagrams

This is the key for the object model diagrams youwill find throughout this book:

Abstract-Class

Instantiable-�Class

Typeinheritance

Instantiation

Association

Aggregation

Composition

1..*

Multiplicity

Createable-�Class

This notation is based on the UML notation, anindustry diagramming standard for object-orientedanalysis and design.

The object model diagrams are an importantsupplement to the information you receive in objectbrowsers. The development environment, VisualBasic or other, lists all of the many classes andmembers, but does not hint at the structure of thoseclasses. These diagrams complete yourunderstanding of the ArcInfo components.

This book uses UML to document the ArcInfosoftware components, ArcObjects, and to illustratecustom data models that you can build.

Classes and objects

There are three types of classes shown in the UMLdiagrams—abstract classes, createable classes, andinstantiable classes.

An abstract class cannot be used to create newobjects, but it is a specification for subclasses. Anexample is that a “line” could be an abstract classfor “primary line” and “secondary line” classes.

A createable class represents objects that you candirectly create using the object declaration syntax inyour development environment. In Visual Basic,this is written with the Dim As New <object> orCreateObject(<object>) syntax.

An instantiable class cannot directly create newobjects, but objects of this class can be created as aproperty of another class or created by functionsfrom another class.

In the Visual Basic object browser, you can inspectall of the ArcInfo createable and instantiableclasses, but not the abstract classes.

Relationships

Among abstract classes, createable classes, andinstantiable classes, there are several types of classrelationships possible.

Associations represent relationships between classes.They have defined multiplicities at both ends.


1..*1..*Owner Land parcel

In this diagram, an owner can own one or manyland parcels and a land parcel can be owned byone or many owners.

A Multiplicity is a constraint on the number ofobjects that can be associated with another object.This is the notation for multiplicities:

1—One and only one. Showing this multiplicity isoptional; if none is shown, “1” is implied. 0..1—Zero or one M..N—From M to N (positive integers)

* or 0..*—From zero to any positive integer 1..*—From one to any positive integer

Type inheritance defines specialized classes thatshare properties and methods with the superclassand have additional properties and methods.

Line

Primary�line

Secondary�line

This diagram shows that a primary line (createableclass) and secondary line (createable class) aretypes of a line (abstract class).

Instantiation specifies that one object from oneclass has a method with which it creates an objectfrom another class.

TransformerPole

A pole object might have a method to create atransformer object.

Aggregation is an asymmetric association in whichan object from one class is considered to be a“whole” and objects from the other class areconsidered “parts.”

Transformer�bank

Transformer3

A transformer bank has exactly three transformers.In this design, transformers can be associated witha transformer bank, but may also exist after thetransformer bank is removed.

Composition is a stronger form of aggregation inwhich objects from the “whole” class control thelifetime of objects from the “part” class.

CrossarmPole1..*

A pole contains one or many crossarms. In thisdesign, a crossarm cannot be recycled when thepole is removed. The pole object controls thelifetime of the crossarm object.

Expressing models with diagram notation

If you are unaccustomed to this type of diagramnotation, practice reading the examples above andconceive of your own examples. Before long, youwill read these diagrams with ease. You will findthat it is worth your effort to understand thisnotation. It describes object models in a conciseand expressive way and will facilitate yourconceptual understanding of the ArcInfo softwarecomponents.

Understanding this notation is also critical if youcreate custom features by extending thegeodatabase data access objects. With ArcCatalog,you can launch a computer-aided softwareengineering (CASE) environment to create customdata models with a visual user interface. Thisinterface is based on manipulating graphicalsymbols from the UML notation.


TECHNOLOGY TRENDS

A geographic information system is at its core adatabase management system enhanced to store,index, and display geographic data.

ArcInfo 8 is a significant release of new GIStechnology that exploits several importanttechnology trends just as they have become readyfor commercial implementation. These trendscollectively realize the vision of GIS as ageographically enabled database.

The timing of ArcInfo 8 is fortuitous as it occursduring the convergence of several criticaldevelopments in software and database technology.The following are the principal trends that shapethe technological framework of ArcInfo 8.

Spatial data and databases

When the coverage data model was firstimplemented, practical considerations led to thespatial component of geographic data beingcontained in binary files with unique identifiers torows in relational database tables that stored featureattributes.

With performance and functional advances indatabase technology, it is now possible andadvantageous to store all spatial data directly withinthe same database tables as attribute data.

The gain from storing spatial data directly withincommercial databases is improved dataadministration, the utilization of data access andmanagement services, and closer integration with theother databases that an organization manages.

Moreover, ArcInfo users can select from any of theindustry-leading relational databases to host theirgeographic databases.

User interface

Applications developed for Microsoft Windows®

have set a new standard for ease of use andconsistency. Users have become accustomed toexpected behaviors for mouse interaction, menus,dialog boxes, and the like. These user interfacestandards have made powerful applicationsaccessible and usable by people who are notcomputer experts.

ArcInfo 8 thoroughly implements the Windowsstandards for user interface and stands as a newmilestone in making GIS software easier to use.

Software component architecture

Modern software is built on software componentarchitectures, examples of which are MicrosoftComponent Object Model (COM), the CommonObject Request Broker Architecture (CORBA), andJava Remote Method Invocation (RMI).

The idea behind components is to divide softwarefunctionality into discrete, independent pieces thatcan be developed, tested, and combined intoprograms. By their design, components can be usedto build any number of applications withoutmodification. This is a high level of software reuse.

The benefit of software component architectures isbetter software quality, better performance, and theability to update software versions without affectingother installed software.

ArcInfo 8 is built on the Microsoft COM architecturebecause it is the most robust and reliablecomponent framework for desktop applications.

Programming environment

Visual programming environments such as VisualBasic have become the norm for applicationdevelopment.

The benefits of using these languages are the largepool of experienced programmers and the richnessof these environments. It is no longer necessary ordesirable to use proprietary macro languages.

ArcInfo 8 uses Visual Basic for Applications (VBA)as its embedded macro language for customizing itsapplications, ArcMap and ArcCatalog. Other COM-compliant languages such as Visual C++ can beused to extend the geodatabase data model.

Trends in summary

The common themes of these technology trends areopen standards and interoperability.

The benefit of implementing these trends is to takeadvantage of technology from other industrysegments, which lets ESRI concentrate its researchand development on core GIS functionality.

23

2 How mapsinform

A map is the principal transmitter ofknowledge for a geographic informationsystem. Maps let people effectivelyrecognize spatial patterns, relationships, andtrends. This chapter discusses the followingtopics:

• The utility of maps

• How maps present information

• The parts of a map

• Presenting geography through layers

• Drawing features with symbols

• Drawing methods for feature layers

• Classifying attribute values

• Displaying thematic and spectral datawith raster layers

• Visualizing surfaces with TIN layersSouth and Central America, Arnold Florentin van Langren, 1596.


People have used maps throughout history. Untilrecently, maps were exclusively printed documents.Drawn on flat sheets of paper or parchment, mapsdepicted objects in the real world—paths,settlements, and natural features.

The practice of cartography evolved to supportdiverse and inventive ways to characterize the manyqualities of the real world. Techniques weredeveloped to depict classifications of features,identifying labels, the shape of the earth’s surface,and the flow of resources or goods.

Many of these practices are manifest in our modernmaps, such as the use of double-line symbols forroads, text label placement, and the application ofthe color blue for bodies of water.

With the widespread adoption of computers andthe development of GIS technology, maps are nowthe printed documents with which we are familiar,as well as interactive visual displays on computers.

GIS systems have enhanced the way people interactwith maps. You can easily define the manner inwhich information is presented and can also selectlocations or objects to initiate a query or analysis.

WHAT MAPS DO

Maps are uniquely capable for sharing knowledgeabout our world in many ways.

Maps identify what is at a location. You can pointto a location on a map and learn the name of theplace or object and any other descriptive attributes.

Maps can locate where you are. If your map hasreal-time input from the Global Positioning System(GPS), you can see where you are, how fast youare traveling, and the direction you are headed.

Maps let you identify distributions, relationships,and trends not otherwise discernible. Ademographer can compare maps of urban areascompiled in the past with present-day maps toguide public policy. An epidemiologist can correlatethe locations of rare disease outbreaks withenvironmental factors to find possible causes.

Maps can integrate data from diverse sources into acommon geographic reference. A municipal

government can merge street maps with maps fromutilities to coordinate construction. An agriculturalscientist can couple images from weather satelliteswith maps of farms and crops to boost productivity.

Maps let you combine and overlay data to solvespatial problems. A state or provincial governmentcan combine many layers of data to find suitablelocations for a waste disposal site.

Maps can find the best path between one place andanother. A package delivery firm can find the mostefficient route for trucks. A public transportationplanner can create optimized bus routes.

Maps can model future events. A utility companycan simulate the impact of a new subdivision anddetermine the necessary system upgrades. Aregional planner can model serious accidents suchas a toxic spill and develop evacuation scenarios.

WHAT MAPS ARE

GIS technology has broadened our view of a map.Instead of a static entity, a map is now a dynamicpresentation of geographic data.

A map is the graphical presentation of geographicdata. To be effective, a map must be visuallycompelling. Principles of graphic design—layout,proportion, balance, symbology, and typography—apply to maps as well as to other types of illustration.

A map is the interface between geographic data andour perception. Maps utilize people’s inherentcognitive abilities to identify spatial patterns andprovide visual cues about the qualities ofgeographic objects and locations.

A map is an abstraction of geographic data. A mapis a view of geography for a particular class of user.A map filters information for intended use—onlyinformation for the intended purpose is displayed.A map simplifies data—some of the complexity andinternal structure of data is hidden. A map addsdescriptive content to data—labels reveal names,categories, types, and other information.

The goal of a data modeler is to design a datastructure that supports the creation of informativeand aesthetic maps. Understanding how mapsinform is the prerequisite to building a data model.

THE UTILITY OF MAPS

Chapter 2 • How maps inform • 25

When you read a map, you observe facts about theshape and position of geographic features, theattribute information associated with geographicfeatures, and the spatial relationships amongfeatures.

HOW MAPS EXPRESS GEOGRAPHICINFORMATION

Geographic features are located at or near thesurface of the earth. They can occur naturally(rivers, vegetation, and peaks), can be constructions(roads, pipelines, and buildings), and can besubdivisions of land (counties, land parcels, andpolitical divisions).

Three primary ways of presenting a geographic areaon a map are as a set of discrete features, as animage or sampled grid, and as a surface.

DISPLAYING DISCRETE FEATURES

Many geographic features have distinct shapes thatcan be portrayed by points, lines, and polygons.

Points represent geographic features too small to bedepicted as lines or areas, such as well locations,telephone poles, and buildings. Points can alsorepresent locations that have no area, such asmountain peaks.

Lines represent geographic features too narrow tobe depicted as areas, such as streets and streams, orslices through a surface, such as elevation contours.

Polygons are closed figures that represent the shapeand location of homogeneous features, such asstates, counties, parcels, soil types, or land-usezones.

DISPLAYING IMAGES AND SAMPLED GRIDS

Much of the information we collect about the earthis in the form of aerial photographs or satelliteimages. These images often form a backdrop toother map data.

Similar in format to images are sampled data grids,which represent a continuous phenomenon such astemperature, rainfall, or elevation.

Images and sampled data grids are called rasters. Araster is comprised of a two-dimensional matrix ofcells, which have attributes that represent qualitiessuch as color, spectral reflectance, or rainfall.

DISPLAYING SURFACES

The shape of the earth’s surface is continuous.Some aspects of a surface can be drawn asfeatures, such as ridges, peaks, and streams. Linesof equal elevation can be drawn as contour lines.

To portray the shape of the earth, you can create asurface display that uses a range of colors tocharacterize sun illumination, elevation, slope, andaspect. Most often, the vertical values represent anelevation, but other attributes such as populationdensity can define a surface as well.

HOW MAPS PRESENT INFORMATION


HOW MAPS PORTRAY ATTRIBUTES

The features on a map have any number ofassociated attribute values. These attributes residewithin the database table for a set of features or canbe accessed through links to other databases.

The most common types of attributes are these:

• A descriptive string gives a feature its name orcharacterizes a category, condition, or type.

• A coded value represents a type of feature. It canbe a numeric value or an abbreviated string.

• A discrete numeric value represents somethingthat is counted, such as the lanes on a road.

• A real numeric value represents continuous datathat is measured or calculated, such as distance,area, or flow.

• An object identifier is rarely displayed, but it isthe key to access attributes in external databases.

There are a variety of techniques for illustratingdescriptive information on a map.

Depicting type attributes

Coded values are used to draw symbols that depicta type of object. Points are drawn with recognizablesymbols for schools, mines, and ports. Lines aredrawn with distinct pen patterns that representcontinuous or intermittent streams. Areas are drawnwith fill patterns that portray any classification.

Illustrating measured attributes

Numeric values can be drawn on a map by varyingthe size of symbols. These values can be integers orreal numbers and can be grouped into classifications.

Drawing classified attributes

Coded values or numeric values can be presentedon a map by using colors. A color can representthe features that share a common value. A colorcan represent a numeric value within a range by ablend from one color to another or a gradation inhue, brightness, or saturation.

Labeling descriptive attributes

Taos Ski Valley EgyptRio Puerco

Descriptive strings can be drawn next to, along, orinside the features they describe.

HOW MAPS EXHIBIT SPATIAL RELATIONSHIPS

When you look at a map, your mind discernsspatial patterns. Many maps are built for purposessuch as identifying business locations, optimizingroutes, and understanding habitats.

Maps visually reveal these spatial relationships:

• Which features connect to others

• Which features are adjacent to others

• Which features are contained within an area

• Which features intersect

• Which features are near others

• The difference in elevation of features

• The relative position among features

Maps in a GIS also support spatial queries thatcreate lists and selections.


ArcInfo and its mapping application, ArcMap,present a model of a digital map that conforms toour experience with traditional maps.

You can print this digital map on a large-formatprinter to high cartographic standards. You caninteract with this digital map on a computer andmodify the thematic display, query features,perform analysis, and edit features. The digital mapis stored as a file with an extension of .mxd and iscalled a map document, or simply, a map.

THE MAP AND ITS ELEMENTS

A map document contains cartographic elementswith which you are already familiar—north arrows,scale bars, neatlines, titles, insets, and legends. Themain elements of a map are organized this way:

• A map has one or more data frames that presentgeographic data.

• Each data frame has one or several mapsurrounds that display a cartographic context.

• The page layout of a map has a number of mapelements that finish the map.

The map’s container of geography

A data frame contains the geographic data on yourmap. A map can have one or several data frames.

A data frame has one or many layers that arestacked on top of each other and span the samegeographic extent. A data frame occupies an extenton the page layout and spans a geographic extent.The ratio between a data frame’s geographic extentand its layout extent is the map scale.

A data frame has a coordinate system that describeshow that part of the world is projected. Thiscoordinate system may be the same as or differentthan the coordinate system of the layers.

The cartographic qualities of a data frame

A data frame can be associated with map surroundsthat present the cartographic context such as scale.

Map surrounds are dynamically linked to a dataframe. When the drawing method is changed, thelegend is updated. When the map scale is changed,scale text is updated and the scale bar is resized.When the map is tilted, the north arrow is rotated.

The finishing graphics of a map

You can add map elements to complete your map.

Map elements include markers, lines, polygons,rectangles, text, and pictures. A picture can be aWindows metafile or bitmap. Map elements have noexplicit association with the data frame.

THE PARTS OF A MAP


A layer is the basic unit of geographic presentationon a map. It shows a set of related geographic datadrawn to a cartographer’s specifications. Someexamples of layers you might create are streams,political boundaries, survey points, and roads.

LAYERS ABSTRACT GEOGRAPHIC DATA

A layer is a reference to a set of geographic data,but it does not contain geographic data. There areseveral advantages to this approach:

• You can create distinct layers on the samegeographic data that visualize different attributesor employ different drawing methods.

• You can edit geographic data, and map layersare updated the next time you display the map.

• Layers are shared across an organization withoutduplicate geographic data. A layer can referencedata at any location accessible on a network.

A layer is stored as a part of a map document or asa separate file on your computer disk with anextension of .lyr. You can think of a layer as acartographic view of geographic data. A layer letsyou assign drawing methods, set scale thresholds,and apply selections to the display.

Drawing many views of geographic data

A layer lets you assign any type of drawing methodto a geographic dataset.

A geographic dataset of the world’s countries might have anumber of attributes such as population, life expectancy,growth rate, and water quality.

However, geographic datasets do not contain theinstructions for drawing the data. You specify themethods for drawing data when you create a layer.

You can create multiple layers for the same dataset.Each of these layers can depict a separate attribute.

These maps show life expectancy, water quality, andpopulation growth in South America.

Drawing selections of features

Some maps show subsets of features in a dataset.When you create a layer, you can select featuresinteractively on the map or specify an attributequery using Structured Query Language (SQL)syntax.

The first map shows all the countries in Europe; the secondshows those countries participating in currency unification.

With selections in a layer, you can draw only thefeatures of interest without having to delete features.

Controlling the map scale of layers

You can draw a map to any map scale, but certainlayers are best drawn within a prescribed scalerange. You can set a scale threshold for a layer andreplace one layer with another at a specified scale.

small scale

large scale

The first map shows a layer with buildings drawn with fillsymbols. The second map shows a layer with the samegeographic dataset, but drawn with marker symbols.

PRESENTING GEOGRAPHY WITH LAYERS


TYPES OF LAYERS

Recall that a geographic area can be presented on amap as a set of discrete features, as images or grids,or as surfaces. Below are some of the types oflayers you can add to a map.

Most layer types are associated with geographicdatasets within geodatabases. Subsequent chaptersof this book contain further information on thesedata objects.

Mapping discrete features

Many geographic objects have a distinct shape.

A feature layer uses a drawing method to presentdescriptive information about a feature class. Afeature class is a homogeneous collection of point,line, or polygon features.

Mapping images and sampled grids

Much of the geographic data that is collected is inthe form of satellite imagery, photographs, or grids.

A raster layer uses a drawing method to presentspectral or descriptive information about a raster. Araster is a matrix of cells with attribute values.

Mapping surfaces

Surfaces represent the shape of the earth.

A TIN layer uses a drawing method to show thez value of a triangulated irregular network (TIN).A TIN is composed of adjacent triangles that sharenodes and edges.


Maps present descriptive information aboutgeographic features using symbols and labels.

Here are some common ways that maps presentdescriptive information about the geographicfeatures they represent:

• Roads are drawn with various widths, patterns,and colors to represent different road classes orother attributes.

• Streams and water bodies are typically drawn inblue to indicate water.

• Special symbols denote specific features, such asrailways and airports.

• Streets are labeled with names.

• Buildings can be labeled with their name orfunction.

Four basic types of symbols are used to presentdescriptive information about features: markersymbols, line symbols, fill symbols, and textsymbols.

Drawing points with marker symbols

You can choose from several types of markersymbols to represent point features on a map.

Character marker symbol

Simple marker symbol

Arrow marker symbol

Picture marker symbol

Multilayer marker symbol90

A character marker symbol is based on a singlecharacter (or glyph) in a TrueType® font. Thesesymbols are drawn with one color.

A simple marker symbol is a predefined simplestroked symbol such as a square or circle optimizedfor rapid screen display.

An arrow marker symbol is based on a singlecharacter in a predefined TrueType font for thepurpose of drawing arrows (or line decorations) atthe ends of cartographic lines.

A picture marker symbol is a bitmap or enhancedmetafile. A bitmap is a standard Windows® raster

image with a file extension of .bmp. An enhancedmetafile is a standard Windows vector drawing witha file extension of .emf. Enhanced metafiles canhave many colors and, because they are based onvector graphics, can be drawn at different sizeswithout visual degradation.

A multilayer marker symbol is a composite symbolthat combines any of the other types of markersymbols. This is ideal for complex symbols that area combination of shapes and text, such as highwayshield symbols. A simple marker symbol can beused as an outline for a multilayer marker symbol.

Drawing lines with line symbols

Linear features on a map can be drawn with one ofthe following line symbols:

Cartographic line symbol

Hash line symbol

Marker line symbol

Multilayer line symbol

A cartographic line symbol is a general-purpose linesymbol with display properties of width, color,parallel offset distance, dash pattern (or template),arrow heads (or line decoration), cap, and join. Capspecifies whether the ends of line symbols aredrawn squared, butted, or rounded. Join specifieswhether corners of lines are square, rounded, orbeveled.

A hash line symbol has short segments that areperpendicular or at any specified angle to the pathof a line. Hash line symbols are usually combinedwith cartographic line symbols within a multilayerline symbol; the customary symbol for railroadtracks is an example of this.

A marker line symbol contains marker symbols in apattern defined by a template. Any type of markercan be placed within a marker line symbol.

A multilayer line symbol is a composite symbol thatcombines any of the other types of line symbols.The example of a railroad track symbol is achievedby combining a cartographic line symbol with ahash line symbol in a multilayer line symbol.

DRAWING FEATURES WITH SYMBOLS


Drawing areas with fill symbols

You can draw areal features with one of these fillsymbols:

Simple fill symbol

Line fill symbol

Marker fill symbol

Gradient fill symbol

Picture fill symbol

Multilayer fill symbol

A simple fill symbol has display properties of color,outline style (null, solid, dashed, and others), andoutline width. A simple fill symbol can also containa number of predefined line fill patterns such ashorizontal hatch or crosshatch. Simple fill symbolscan be hollow; you can draw areal features byoutline only.

A line fill symbol has the properties of a simple fillsymbol, but you can specify a richer type of line fillpattern that can incorporate any line symbol at anyangle and separation.

A marker fill symbol is drawn either as a grid ofmarker symbols that can be arbitrarily spaced androtated, or as a random distribution of markersymbols with a specified average horizontal andvertical separation.

A gradient fill symbol is drawn as a blend of twocolors, transitioning from one to another. There arefour types of gradient fill:

• A linear gradient blends colors in one direction,from top to bottom, left to right, or at any angle.

• A radial gradient blends colors in a circularpattern from the center point outward to theouter part of the area.

• A rectangular gradient blends colors from thecenter outward in a rectangular pattern.

• A buffered gradient blends colors inward fromthe perimeter of an area. A percentage valuelimits how far the gradient progresses inwardfrom the perimeter. This is ideal for thecartographic convention of drawing oceanshorelines.

A picture fill symbol is comprised of bitmaps orenhanced metafiles. The pictures are drawncontiguously or with a fixed spacing.

A multilayer fill symbol is a composite symbol thatcombines any of the other types of fill symbols.


A feature layer is a reference to a feature class andhas an associated drawing method (or renderer). Youcan choose any string or numeric attribute of yourfeature layer and visualize it in a variety of ways.

You will find that the type of attribute you areinterested in visualizing guides your selection of adrawing method. Numeric data might be bestpresented with symbols that change size or coloraccording to the attribute value. Attributes thatdescribe a type of feature might be best drawn withsymbols that match each unique value.

The following sections outline the drawing methodsavailable for feature layers.

DRAWING FEATURES

The simplest way to draw a feature layer is to drawall the features with the same symbol.

With this drawing method, all features are drawnwith a symbol that follows a cartographicconvention. Well heads could be drawn with asquare marker, streams could be drawn with bluelines, and buildings could be drawn with simpleyellow fill symbols with a black outline.

This map draws all countries with the same fill symbol.

This drawing method is suitable for feature layersthat represent a fairly homogeneous set ofgeographic features. It is also used for a simpledisplay of feature layers that are behind other layersof greater interest.

This method is also best for drawing features simplyso that spatial distribution patterns can be visuallyrecognized. If your map contains point features forpotential customers, this drawing method can helpyou discern spatial clusters for geographicallytargeted marketing.

This drawing method is called the simple renderer inthe ArcInfo object model.

DRAWING CATEGORIES OF FEATURES

The feature’s attribute of interest can be drawn bycreating categories. The following are thetechniques used to symbolize by category.

Drawing categories by unique field values

An attribute in a feature layer sometimes representsan important subdivision of the feature type. Thisattribute can describe a category of a feature, suchas a land-use type or type of road. It can alsocharacterize a relation between the feature and alarger entity, such as a province or state and thecountry to which the feature belongs.

12 Pole

1 Pedestal

61 Transformer

12 Residential

46 Lake

69 Park

1 Railroad

30 Highway

94 Canal

Value

Line symbol

ValueLabel

Label

Label Value

This drawing method lets you assign a uniquesymbol to each unique value of the attribute. Anelectric device layer can contain a type attributerepresenting poles, pedestals, and transformers. Atransportation layer can have a type attribute forrailroads, highways, and canals. A land-use layercan have a classification attribute designatingresidential, lake, or park status.

DRAWING FEATURE LAYERS


Roadway

Commercial

Residential

Park

Industrial

Land use

This map shows areas drawn with a unique symbol for eachtype of vegetation.

With this type of drawing method, you can surmisehow different types of features are located withrespect to each other and their relative frequencyand distribution.

This drawing method is called the unique valuerenderer in the ArcInfo object model.

Drawing categories by unique combined fieldvalues

You can elect to draw categories of the uniquecombinations of up to three field values. In thisplanning map, you can see combined values ofland use and historic district zones.

Use this drawing method with care. It is not difficultfor the number of unique combined values tobecome too large to visually discern theclassifications you want to differentiate.

Roadway

Commercial, None

Residential, None

Park, None

Industrial, None

Land use and historic district

Commercial, Old Town

Residential, Old Town

Park, Old Town

Industrial, Old Town

This map shows unique combinations of field values.

Drawing categories by symbol names in a field

You can draw features by using symbol names in afield. This field contains symbol names as textvalues such as “Primary road,” “Industrial zone,” or“Survey marker.” This field can have any name.

This drawing method is the easiest way to ensurethat symbols are drawn exactly the same way indifferent maps throughout an organization.

Name Type Symbol

dirt road

residential road

highway

feature layer table style

residential road

dirt road

highway

highway

Another advantage of this drawing method is thatan organization can implement parallel styles toproduce distinct sets of maps. For example, youcan create a set of styles with symbols intended fordifferent map scales. At one map scale, a road canbe drawn as a simple line, and at another scale, thatroad can be drawn with double lines. This drawingmethod makes it easy to switch how symbols aredrawn for different map products. Features drawnwith this method appear the same as features drawnby unique values.

DRAWING QUANTITIES OF FEATURES

Numeric fields can store values that are numericallyordered and can represent counted or continuousvalues. The following are the drawing methods tovisualize the quantities of features.

Drawing quantities by color

An effective way to display a numeric attribute is topresent an attribute with a set of graduated colors.This attribute, called a value field, can benormalized by another field. This means one valueis divided by another.

The value field is subdivided into a set of classes.You have several options for classification, and thisis discussed in some detail in the next chaptertopic. Classification is a statistical process forsubdividing a collection of values.


The graduated colors are set by selecting a colorramp, which is a transition from one color toanother. A color ramp can have multiple parts—abeginning color can blend to an intermediate colorand then to a final color.

2.2–6.3

6.3–9.2

9.2–13.4

30–49

50–79

80–99

8.8–9.8

9.8–23.2

23.2–54.1

Marker symbol

Value range

Line symbol

Value range

Fill symbol

Value range

This drawing method is an effective presentation fornumeric data that represents a continuous attributesuch as elevation, temperature, or amount of aresource. It is especially effective for drawing areas.

Color ramps are used to apply cartographicpractices. For example, bathymetric maps are drawnwith light blue for shallow waters to dark blue fordeep waters.

This map shows population density by drawing a populationattribute normalized by an area attribute.

This drawing method is called the class breaksrenderer in the ArcInfo object model.

Drawing quantities by graduated symbols

Another effective way to visually present a numericattribute is to vary the size of a symbol. Again, youspecify a value field, an optional normalizationfield, and a classification. The range of values issubdivided into the number of classes set in theclassification.

Instead of a color ramp, you select a base symboland a range of symbol sizes.

0.2–0.4

1.2–1.41.0–1.2

0.8–1.0

0.6–0.8

0.4–0.6

Line symbol

Value range

Marker symbol

1–2

10–14

7–9

3–6

Value range

Marker symbol

inside polygon

0–100

201–300

101–200

Value range

This drawing method is suitable for numeric datathat represents a rank or progression of values.Some care should be taken in selecting the range ofsymbol sizes so that features do not overlapexcessively in dense areas.

This method draws the larger symbols first and thesmaller symbols afterward. This is so that featureswith large values do not obscure features withsmaller values. You can choose contrasting colorsfor the outline and body of the symbol to makeoverlapping features stand out.

An interesting behavior of this drawing method isthat value ranges for areas are drawn with markersymbols instead of fill symbols. That is because thesize of an area is predetermined by its shape, anddrawing a marker symbol at its centroid point is analternative to drawing a graduated symbol.

This map draws cities with marker symbols graduated bypopulation value against a layer with administrative areas.

This drawing method is called the graduated symbolrenderer in the ArcInfo object model.


Drawing quantities by proportional symbols

This drawing method is similar to drawing featureswith graduated symbols, except that there is noclassification of values and the size of each symbolvaries in exact proportion to the attribute value.

You specify a value field, an optional normalizationfield, and the units for display. You can also specifywhether the attribute value varies the symbol’s sizeby area or radius.

1

10

Marker symbol

Value

0.1

1.2

Line symbol

Value

45.1

97.3

Marker symbol

inside polygon

Value

This drawing method is suitable when you want todraw a continuous gradation of symbol sizes.

As with the graduated symbol drawing method,symbols can overlap each other. Therefore, largersymbols are drawn first, followed by smallersymbols, and the values for area features are drawnwith marker symbols.

This map shows countries that are sized in exact proportion totheir population.

This drawing method is called the proportionalsymbol renderer in the ArcInfo object model.

DRAWING MULTIPLE ATTRIBUTES

On occasion, you will want to symbolize featuresby multiple attribute values.

This drawing method lets you effectively use tworenderers at once on a feature layer. This method iscalled the bi-unique value renderer in the ArcInfoobject model.

This map shows the countries of Europe drawn with twodistinct attributes symbolized by color coding a unique valueon the polygons and a quantitative value on the point symbols.


Two of the drawing methods for feature layers—drawing features with graduated colors and drawingfeatures with graduated symbols—are based on aclassification of attribute values. A drawing methodfor raster layers—drawing cells by graduatedcolors—also uses a classification of attribute values.

CLASSIFICATION METHODS

A classification method is applied toward a groupof attribute values and subdivides them accordingto the desired criteria. These can be equalsubdivisions of the range of attribute values, equalcounts of features within each class, or anothercriterion.

Each subdivision of the group of attribute values isknown as a class. Each class has a lower and uppernumeric range limit. For each of these classescalculated by any of the classification methods, youcan override a class range limit and set another.

Classifying values by natural groupings

The natural breaks classification uses a statisticalformula to determine natural clusters of attributevalues. The function of the formula, known asJenk’s method, is to minimize the variance within aclass and to maximize the variance between classes.

The graduated symbol and graduated color drawingmethods apply this classification method by default.

Features, sorted by attribute value

50

60

40

30

2010 0–30

38–52

60–68

Attributevalue

Attr

ibut

e va

lue

The natural breaks classification is well suited touneven distributions of attributes. Distinct naturalgroupings of attributes can be isolated andhighlighted.

Classifying values by defined intervals

The defined interval classification divides a set ofattribute values into classes that are divided byprecise numeric increments, such as 10, 100, or 500.


50

60

40

30

2010 0–25

26–50

51–75

Attributevalue

Attr

ibut

e va

lue

This classification works well for values that peopleare accustomed to seeing in rounded numbers,such as age distribution, income level, or elevationranges. The disadvantage is that some of theclasses, particularly the first and last, may contain adisproportionate number of feature values.

Classifying values by equal intervals

The equal interval classification takes the range ofvalues and subdivides them into ranges of equalvalue intervals.

A value range from 21 to 69 with three classeswould be subdivided into range spans of 16 units,21–36, 37–52, and 53–69.


50

60

40

30

2010 21–36

37–52

53–69

Attributevalue

16

16

16

Attr

ibut

e va

lue

This classification emphasizes how feature valuesfall within uniform ranges of attribute values. Inpractice, it is similar to defined intervals, but has theadvantage that the lowest and highest class span thesame range of values as the rest of the classes.

An example of the application of this classificationis a map that depicts homes for sale divided intoequal ranges of purchase costs.

Classifying features by quantiles

The quantile classification creates classes withequal numbers of features. If a feature layer has12 features, three classes would each represent fourfeatures.

CLASSIFYING ATTRIBUTE VALUES



50

60

40

30

2010 0–41

42–49

51–68

Attributevalue

4 4 4A

ttrib

ute

valu

e

This classification is particularly effective for rankedvalues. A company can measure sales performanceof business locations and draw the respectivebusinesses in their rank of sales performance. Thisclassification yields visually attractive maps becauseall of the classes drawn have the same number offeatures.

However, this classification might obscure thenatural distribution of attribute values; clusters ofattribute values may be split or combined with othervalues. This classification is best applied to attributevalues that have a generally linear distribution.

If you have an even number of classes, the valuedelimiting the middle classes is the same as themedian of a statistical sampling.

Classifying features by standard deviations

The standard deviation classification creates aneven number of classes that represent whole orfractional deviations from a mean value.

First, the mean (or average) of all the attributevalues is calculated. Then a statistical formulacalculates a standard deviation.

You can specify the number of classes and whetherthey span one whole standard deviation, one-halfstandard deviation, one-third standard deviation, orone-quarter standard deviation. The classes at thelow and high end extend to the minimum andmaximum values.

32.3–46.1

46.1–58.8

58.8–69

Attributevalue

Mean is 46.1, one standard deviation is 13.8

0–32.3


50

60

40

30

2010

mean +1 s.d.-1 s.d.

Attr

ibut

e va

lue

This classification is intended for generallysymmetric distributions of attribute values that havea broad peak near the mean with the density ofvalues diminishing away from the peak.

An example of a suitable map for this classificationcould be population density or accident rates. Youwould expect these values to have their greatestdata density near a mean value and that values thatvary significantly are scarce. The classic shape ofthis type of distribution is the bell curve.

Normalizing attribute values

Sometimes, a classification is best applied not to asingle attribute, but to one attribute normalized byanother. Normalization is simply dividing oneattribute value by another.

An example of a normalized attribute would beaccident rate. The data might contain accumulatedvalues for sections of highways, but for accidentdata to be meaningful, the number of accidentsshould be normalized (or divided) by the length ofeach highway segment.

Excluding attribute values

Some data contains erroneous or null values, oryou might want to examine only a certain range ofvalues.

Erroneous data might be flagged as values outside areasonable range. For example, percentage valuesmight always be expected to be from zero to 100.Smaller and larger values would be excluded.

Attribute values that represent a ranking of featuresmay have a coded value such as –99 that representsa null or unknown value. This value should beexcluded.


Much of the most readily available geographic datais in the form of rasters. A raster is a regularlyspaced matrix of cells that may have associatedattribute information.

A cell can represent either continuous data such aselevation and pollution concentration, or spatiallydiscrete data such as land use or vegetation type.

A raster can have a single band or multiple bands. Aband is like a layer that represents different valuesfor each cell. A common example of this value islight reflectance at a part of the spectrum.

A raster layer is a reference to a raster with aspecified drawing method. The same raster can bedrawn with several raster layers, each with adrawing method to emphasize a particular attributeor classification.

Chapter 9, “Cell-based modeling with rasters,”contains more information on the raster datastructure, its advantages for modeling, and the typeof analysis possible with rasters.

TYPES OF DATA IN A RASTER

A raster contains one of three types of information:thematic data, spectral data, or pictures.

Thematic data in a raster

A raster can represent a particular phenomena,such as fire, chemical concentration, slope, orelevation. These are typically stored as single-bandrasters and often have associated attribute tables.

This raster layer draws the aspect of a terrain.

Aspect is the direction toward which a section ofsurface is pointing. In this raster, red denotes slopesfacing north and yellow shows slopes facing south.

Spectral data in a raster

The most common use of a raster is to presentimages of the earth acquired through aerialphotography or satellite imagery. Specializedcameras can capture the reflectance of light atseveral or many parts of the spectrum.

This raster layer shows a multiband spectral raster capturedfrom a satellite imaging system.

In the hands of an imaging scientist, these imagescan be compared with known spectral signatures ofrocks or plants to reveal geological or vegetativestructures.

Picture data in a raster

A raster can contain pictures such as scanned mapsor building photographs. This type of raster can besingle or multiband.

This raster layer shows a picture of a house.

Building pictures can be useful in a property assetapplication.

DISPLAYING THEMATIC,SPECTRAL,AND PICTURE DATA


DRAWING METHODS FOR RASTER LAYERS

A raster layer is a reference to a raster and has anassociated drawing method. You can select anyattribute of your raster layer and visualize it in avariety of ways.

There are two broad types of raster: single-bandand multiband. Some rasters are imaged and someare sampled from other data. The drawing methodsfor raster layers are described below.

Drawing cells by unique value

A raster can be optionally associated with a tablecontaining attributes for each cell. These cellattributes can describe spatially discrete (orthematic) data such as landuse, soils type, andproperty ownership.

Water

Agricultural

Residential

Industrial

Attribute value

This drawing method is suitable when an attributeexists that already describes a category, type, orclassification. The attributes can be descriptive ornumeric. Generally, the number of unique valuesshould not exceed 25. Otherwise, it becomesdifficult to distinguish between classes.

This raster layer uses unique vegetation codes to draw thedistribution of plants for a region.

If a raster has one-bit data, this drawing methodcan be used with 0 for black and 1 for white tomake a monochrome drawing.

Drawing cells classified by graduated colors

Some cell attributes represent a range of numericvalues that contain thematic information, such aselevation, slope, pollution contaminants, orpopulation density.

1–4

5–8

9–12

13–15

Value

This drawing method lets you define a classificationin the same manner as the graduated symbol andgraduated color drawing methods for feature layers.You can normalize and exclude attribute values. Ifa raster has multiple bands, you will select one ofthose bands for this drawing method.

Once a classification has been made, you canselect a color ramp showing each class in adifferent color. The color ramp you select shouldprovide visual cues that your perception is alreadyaccustomed to.

For example, denser concentrations should usebold colors and lighter concentrations should bepale. Temperatures should be blue for cold to redfor hot. An elevation map can have two ranges ofcolor, one for elevations above sea level andanother for bathymetric elevations.

This raster layer shows discrete elevation ranges displayedwith a color ramp.


Drawing cells stretched with graduated colors

Many rasters have continuous data that represents aspectral value, or a calculated value, such as sunillumination angle.

High

Medium

Low

Value

This drawing method is traditionally used for single-band continuous data with a large set of values. Itcreates high-quality displays of continuousphenomena by rendering them with a continuousgradation of colors.

This raster layer shows a raster with sun-illumination anglevalues stretched from white to black. This type of map is calleda hillshade map.

For a raster band, you select the type of stretch. Astretch is like a classification, but it describes therate of change of continuous values.

Some of the available types of stretches includestandard deviation, histogram equalize, andminimum–maximum. The stretch calculates andassociates high, medium, and low numeric valueson a color ramp. Areas of no data can be drawnwith a separate color.

Drawing cells using a red-green-blue composite

Rasters that are created for color display are oftencreated with three bands, one each for red, green,and blue.

The types of data collected with these three-bandrasters can be satellite imagery, scannedphotographs, or any type of picture.

Redband

Greenband

Blueband

Red-green-bluecomposite

0

255

Attribute valuesrange from 0 to 255

in each band

This drawing method uses a collective stretch forthe three bands. Areas of no data are drawn with aspecified color.

This raster layer shows a red-green-blue color compositesatellite image of an urban area adjacent to mountains.


A triangulated irregular network (TIN) is an efficientrepresentation of a surface. Rasters are also usedfor surface modeling, but TINs have the advantageof varying the data density where the surfacechanges sharply. Areas where the surface is smoothrequire few points, while areas where the surface isrough require more points.

THE ELEMENTS OF A TIN

A TIN is made from points, each of which have acontinuous real value that usually representselevation. Other types of surface values might bechemical concentration, groundwater level, orprecipitation amount.

A triangulation is calculated from these data pointsand represents a continuous, three-dimensionalsurface. A triangulation is a nonoverlapping set oftriangles, or faces, that completely fills a prescribedarea.

Because TINs represent a surface with vectorfeatures (points, lines, and faces), they can preciselymodel discontinuities in the surface with breaklines.Examples of breaklines are streams, ridges, androadways, where the surface slope changes sharply.

One limitation of a TIN is that it cannot representthe rare occurrences of negative slope, such as onvertical cliffs, overhangs, and caves.

z

y

x

node edge

face

normal

This is a perspective view of one face in three dimensions.

A face defines a plane, a slope, and a slopedirection. The normal to a face is the perpendicularvector and it is used for calculations such as sunillumination, aspect, and slope.

Chapter 10, “Surface modeling with TINs,” has moreinformation on the data structure of TINs, data-sampling considerations, the analytic possibilities

with TIN data models, and a comparison of TINsand rasters for representing surfaces.

DRAWING METHODS FOR TIN LAYERS

A TIN layer is a reference to a TIN and anassociation to one or more TIN drawing methods(or renderers).

You can draw a TIN layer with one or manydrawing methods (renderers) that display TINelements (points, lines, and faces) or visualizequalities of a surface such as elevation, slope, andaspect.

Unlike the drawing methods for rasters and features,a TIN layer allows you to select many drawingmethods instead of one. This lets you drawdifferent data elements at once, such as breaklinesdrawn on top of faces that are colored by elevation.

The following sections describe the drawingmethods for TIN layers.

Drawing TIN elements

You can draw the points, lines, and faces in a TIN.

Normally, you would not draw the points and linesin a map presentation, but this option can be usefulfor inspecting or troubleshooting the pointdistribution that makes up your TIN.

VISUALIZING SURFACES WITH TIN LAYERS


Drawing TIN faces with hillshading

All of the drawing methods for faces give you theoption of hillshading, a technique for shading facesto produce realistic views of a terrain.

z

y

x

face

normalsun

illumination angle

Hillshading works by taking the position of the sunin the sky (which you can control) and calculatingthe angle between the direction to the sun and thenormal to a face.

This angle is used to apply shading on faces thatsimulates light reflectance off a surface. Thebrightness of reflected light is proportional to thecosine of the angle between the surface normal andthe vector to the light source.

Hillshading creates a realistic three-dimensionalimage from a two-dimensional display.

This is a TIN layer with faces drawn with hillshading. The sunis in the northwest at an angle of 30 degrees above the horizon.

Drawing elevation with a graduated color ramp

You can draw the faces in a TIN with colors thatshow the range of elevations.

z

y

x

950 m

960 m

This drawing method interpolates contour lines foreach face. A face can have zero, one, or severalcontour lines that cross it. Each zone between thespecified contour interval is drawn with a colorfrom the color ramp.

This TIN layer is drawn with elevations rendered on agraduated color ramp and with linear interpolation.Hillshading is also applied.


Drawing aspect with a graduated color ramp

You can draw the cardinal direction, or aspect, thateach face in a TIN is pointing toward.

z

y

x

face

normal

aspect

Aspect is the direction on a compass that thenormal of a face is pointing to when projected onthe plane of the earth. Aspect is measured indegrees. North is 0 degrees, east is 90 degrees,south is 180 degrees, and west is 270 degrees.

This TIN layer is drawn with aspect rendered on a graduatedcolor ramp. Hillshading is applied.

Drawing slope with a graduated color ramp

You can draw the slope of each face on a surface.This lets you visualize the steep areas of a terrain.

z

y

x

normal

slope angle

Slope is calculated for each face as the anglebetween the normal and the plane of the earth. Acolor ramp is applied to angles between 0 degreesand 90 degrees.

This TIN layer is drawn with slope rendered on a graduatedcolor ramp. Hillshading is applied.

45

3 GIS datarepresentations

GIS is a technology that supports the scienceof geography. The examination of natural orbuilt environments begins with considerationof the basic ways to represent features in theworld.

These are this chapter’s topics:

• The fundamentals of a GIS

• The diverse application of GIS

• Three representations of the world

• Modeling surfaces

• Modeling imaged or sampled data

• Modeling discrete features

• Comparing spatial data representationsOval World Map, Benedetto Bordone, 1528


THE FUNDAMENTALS OF A GIS

This chapter discusses the fundamental concepts youneed to understand in order to build superior datamodels with ArcInfo.

First, a geographic information system (GIS) will bedefined. You will review the parts of a GIS, examinehow GIS extends a database, and explore thediversity of GIS applications.

Next, you will review some basic concepts ofmodeling geographic data. You will learn someoptions for modeling continuous surfaces, discretefeatures, and imagery. Sometimes, there is more thanone reasonable choice for a data model.

THE PARTS OF A GIS

GISSoftware

People

Hardware

Analysis

Data

A geographic information system is the combinationof skilled persons, spatial and descriptive data,analytic methods, and computer software andhardware—all organized to automate, manage, anddeliver information through geographic presentation.

People who build and use GIS

When you design a data model, build a softwareapplication, or write user documentation, it isimportant to be clear on the type of user your workis directed toward.

These are the primary roles that people play in a GIS:

A map user is the end consumer of a GIS. This personlooks at maps created for a general or specificpurpose. All members of the public are map users.

A map builder uses map layers from several sourcesand adds data to make a custom map.

A map publisher prints maps. This person isdedicated to high-quality cartographic output.

An analyst solves geographic problems, such aschemical dispersion, finding the best route, and sitelocation.

A data builder inputs geographic data with severaltechniques—editing, converting, and data access.

A database administrator manages GIS databasesand ensures that the GIS operates smoothly.

A database designer builds logical data models andimplements physical database designs.

A developer customizes GIS software to serve thespecific needs of an industry.

Data sources for GIS

A GIS processes any data that has a spatialcomponent. This information is quite diverse—itcan be aerial photographs or satellite imagery, acollection of terrain contours, digital maps of thebuilt environment, or legal records of landownership.

Geographic data can also reside in some unexpectedplaces—any company that keeps a database of itscustomers has geographic data. A GIS can calculatethe location of any place on earth from a postaladdress.

Chapter 3 • GIS data representations • 47

Procedures and analysis

The specialists that operate a GIS employ functions,procedures, and judgment. This collective humanexperience is an indispensible component of a GIS.

Some examples of analytic functions are:

• A science applied in a spatial context, such ashydrology, meteorology, or epidemiology

• Quality assurance procedures to ensure that thedata is accurate, consistent, and correct

• Algorithms that solve spatial queries on linearnetworks or integrated polygon topology

• The knowledge to apply cartographic designprinciples for excellent map presentation

Computer hardware

Computers come in all sizes, from palm tomainframe. You can purchase GIS software fornearly every type of computer.

With the improvements in network bandwidth, aclient–server or n-tier architecture is the preferredconfiguration for enterprise-scale GIS.

The Internet is joining computers into a globalnetwork and is an important way to access data.Another trend is the increasing use of the GlobalPositioning System (GPS) to locate people in real time.

GIS software: a geographic database

The key idea to grasp about GIS software is that it is,in fact, a geographic database management system.Geodatabases are implemented directly oncommercial relational or object-relational databasemanagement systems.

The reason for this is to leverage the capabilities ofcommercial database software, which include databackup, table definition, transaction management,and system administration tools. A GIS extends arelational database so that it can efficiently storegeographic data, produce maps, and perform spatialanalytic tasks.

data backup

schema definitiontransaction support

system administration

Relationaldatabase

Geographicinformationsystemfeature geometries

spatial index

spatial operators

topology

map rendererscustom objects

security

report generation

versioning

spatial reference

Some of the functions that GIS software adds to arelational database management system are:

• The ability to store the geometric shapes offeatures directly in a database column.

• A framework to define map layers on data andspecify drawing methods; these can be drawnbased on attribute values.

• An infrastructure to support the creation of simpleand sophisticated maps. Many common map-making tasks are simplified.

• The creation and storage of topologic relation-ships that exist among features, such as networkconnectivity and integrated polygon topology.

• A spatial index spanning two dimensions forrapid retrieval of geographic features.

• A set of operators for determining geographicrelationships such as proximity, adjacency,overlay, and spatial comparison.

• Many tools to support spatial queries such asnetwork tracing and polygon overlay analysis.

• A work-flow system that allows the editing ofgeographic data by many users and managesversions.

You can think of a GIS as a spatially enableddatabase management system. This architecture givesyou the best of commercial database technology andsophisticated geographic software.


THE DIVERSE APPLICATIONS OF GIS

GIS is being applied in remarkable ways. Tounderstand GIS and see why it matters, it is useful tosurvey the diverse range of GIS applications.

These are a few descriptions of applications takenfrom papers submitted at the ESRI user conference.

Agriculture

Satellite images of Brazil showing land useare combined with models of El Niño

weather oscillation to predict agricultural effects.

GPS (Global Positioning System) receivers are beingapplied in real time with portable GIS software toaccurately apply chemicals for agriculturalproduction.

In the San Joaquin Valley of California, GIS is usedto model nonpoint sources of pollution. The mapsproduced provide a visual display of soil salinity.

Business geographics

One company used GIS to evaluate howthe pending relocation of its corporate

office would affect employees’ commute to work.

A small company in Quebec facing competitivepressures used GIS to mine its customer databases toidentify clusters of customers, enhance productivityof mail promotions, and improve client retention.

A foundation in San Francisco uses GIS to assistsmall businesses in finding commercial space withdesirable business, economic, demographic, andtransportation attributes.

Defense and intelligence

The U.S. Air Force uses GIS technology tomanage, maintain, and visualize millions

of climatological records.

The Swedish armed forces have done extensivework on sophisticated symbolizing of military andcivilian facilities to improve military planning.

The Canadian Army has customized GIS software tointegrate it with a land force command system.

Ecology and conservation

Colombia is building a GIS database toprioritize which lands should be set aside

for the national park service.

In Kenya, a GIS revealed that large mammals in thesavanna dispersed during the wet season andconcentrated in a basin during the dry season.Understanding seasonal migration patterns isimportant in managing water access for wildlife andlivestock.

GIS is being applied on California’s Santa CatalinaIsland to evaluate the ecological costs and benefitsof dirt roads. Roads pose an environmental dilemmabecause they provide access for ecologicalmanagement but also interrupt the ecologicallandscape.

Electric and gas (AM/FM)

Beirut is analyzing its power circuits tominimize losses and to improve voltage

levels. GIS is modeling scenarios of deviceplacement for optimal electrical benefit.

Public Service of New Mexico is using GIS tomanage the construction, operation, andmaintenance of 2,500 miles of power transmission.A prime concern is preventing environmentallydamaging activities.

The Danish Energy Agency is building a database onthe energy usage of every building in that country.This information will be used for planning energyplants and designing distribution systems.

Emergency management and public safety

In 1997, the Cassini spacecraft waslaunched to explore Saturn. GIS was used

to evaluate the risk of an accident with theplutonium generators on board.

The Italian National Seismic Survey is building anintegrated information system to produce real-timetabular reports and operational maps in the event ofa major earthquake.


Environmental management

In Korea, land zoning in national parksis being analyzed with the criteria of

scenic quality, elevation, slope, and natural state. Itwas found that some parks were not correctly zoned.

A large dam is being constructed in Turkey. GIS isbeing used for a complete evaluation of its effects onirrigation, hydropower, health, mining, education,tourism, and telecommunications.

In Bavaria, ecological balance models are combinedwith GIS software to provide tools for environmentalmanagement. This information is disseminatedthrough the Internet.

Federal government systems

The Tennessee Valley Authority has built aland information system to help administer

land records, natural and cultural resources, land-useplanning, and compliance with laws and executiveorders.

The U.S. National Oceanic and AtmosphericAdministration is building a tool to collect metadatasuch as bounding coordinates, map projections, andattribution information.

Forestry

The construction and use of roads in aforested basin can contribute significantly

to sediment deposition. A forestry company isbuilding a road sediment model to establish amaintenance plan.

The U.S. Fish and Wildlife Service has establishedguidelines for managing forests where the red-cockaded woodpecker, an endangered species, isfound. GIS is used to calculate colony areas andforaging zones.

Health care

The State of California is mandating thatcounty governments address cultural and

ethnic issues for outpatient health care. GIS is beingused to present geographic, socioeconomic,demographic, and health care utilization data.

A university researcher is using GIS to analyze theepidemiology of rare diseases and estimating anindividual’s exposure to environmental risk factors.

In Colorado, the percentage of low-birthweightbabies exceeds the national average. GIS is beingused to examine factors such as age, race, education,elevation, and access to public health programs.

Education

An educational agency is using GIS tohelp students discover geography and

foster critical thinking and inquiry.

A high school is incorporating GIS in its curriculumto teach students a “sense of place” by showingthem how their personal actions have relevance on aglobal scale.

Mining and geosciences

GIS is used in West Virginia to monitoracid mine drainage on surface waters.

Elevations, hydrology, mined areas, and waterquality data are combined.

A mining services company is using GIS to createthree-dimensional databases for nuclear wasterepositories, mineral exploration programs, andgroundwater monitoring purposes.

Oceanography, coastal zone, marine resources

The U.S. Naval Oceanographic Office isusing remotely sensed sea temperature

data to study oceanic fronts and eddies.

In Washington state, a GIS is mapping the currentshoreline, calculating change rates, and projectingshoreline erosion hazards.

Real estate

Habitat for Humanity, an organizationbuilding houses for low-income families,

uses GIS to analyze a proposed subdivision andcreate a plan that preserves most of the existingtrees.

A realty company is using GIS for site selection formultisite users. Factors considered are access,visibility, zoning, and entitlement processes.


Remote sensing and imagery

A digital imagery company is usinggeoreferenced airborne sensors to create

real-time spatial data. Images are sent to groundstations and are fused, reformatted, and subjected toautomatic feature extraction.

State and local government

In Qatar, television cameras are beinginserted into water and sewer networks to

create video records of pipe conditions. Theseimages are integrated within a GIS and giveoperators information for maintenance.

The new Denver International Airport is located in arural area. GIS is being applied to develop scenariosof land-use patterns over the next five, 10 and15 years.

In the Ukraine, political changes have ushered in awave of land reform. Lack of accurate records hashampered the creation of an accurate cadastre, so anew land registration system is being developedbased on high-resolution satellite imagery andinnovative software techniques.

Telecommunications

In Colombia, the fiber-optic trunk networkis being captured in a GIS database with a

representation of each of the network’s elementfeatures.

In Indonesia, GIS is employed to manage radiotelephony by studying radio station placement,the demographics of a customer area, and themaintenance of equipment.

A telecommunications consulting firm is using dataon land use and land cover to predict signalattenuation for wireless communication systems.

Transportation

In Korea, a GIS monitors real-time trafficconditions to mitigate traffic bottlenecks on

freeways.

The State of Georgia applies GIS technology tomanage roadway pavement. A study was made ofroad segment ratings based on load cracking.

Water distribution and resources

Population growth and agriculturalexpansion in Egypt are placing demandson water management. A government

ministry is building a system to manage the NileRiver channel, canals, drains, and pumps.

In Florida, a hydraulic computer model is used toreduce sanitary sewer overflows. When majorrainstorms come, satellite imagery is used to estimaterainfalls and assist in the operation of sewer pumpstations.

In Canada, a hydrodynamic/pollutant transportmodel has been built to simulate the effects ofmultiple pollution sources under different conditions.

SUMMARY OF GIS APPLICATIONS

These applications prove the diversity of GISsolutions. It is always surprising to discover howwidely ranging the uses of GIS technology are.Common characteristics throughout theseapplications include:

• Frequently, GIS is integrated with otherapplications to perform geographic and scientificanalysis. It is important that GIS data bestructured and stored in a way that allows fordistributed data access.

• An open data architecture has considerableimportance in facilitating the integration ofgeographic data with other data, such as real-timedata, imagery, and corporate databases.

• While printed maps are still the most commonpresentation of geographic data, Internet mapaccess and dynamic map applications arebecoming increasingly important for decisionmaking. Interactive access invites moresophisticated data models to support rich queriesand analysis.

• Selecting the right data structure is important toenable the kind of analysis you wish to perform.These applications illustrate many skillfulapplications of modeling the world as acontinuous surface, as a raster grid, or as sets ofdiscrete features in vector format.


THREE REPRESENTATIONS OF THE WORLD

The applications just reviewed show patterns ofhistoric use, capture the current state of the naturaland built environments, and predict changes in theworld based on weather, human activity, orgeophysical events. In every application, a decisionwas made concerning applying physical datasets toserve logical data models.

DATA REPRESENTATION MODELS

With a GIS, you can model data in three basic ways:as a collection of discrete features in vector format,as a grid of cells with spectral or attribute data, or asa set of triangulated points modeling a surface.

Modeling with vector data

Vector data represents features as points, lines, andpolygons and is best applied to discrete objects withdefined shapes and boundaries.

Features have a precise shape and position,attributes and metadata, and useful behavior.

Modeling with raster data

Raster data represents imaged or continuous data.Each cell (or pixel) in a raster is a measured quantity.

The most common source for a raster dataset is asatellite image or aerial photograph. A raster datasetcan also be a photograph of a feature, such as abuilding.

Raster datasets excel in storing and working withcontinuous data, such as elevation, water table,pollution concentration, and ambient noise level.

Modeling with triangulated data

A TIN is a useful and efficient way to capture thesurface of a piece of land.

TINs support perspective views. You can drape aphotographic image on top of a TIN for aphotorealistic terrain display. TINs are particularlyuseful for modeling watersheds, visibility, line-of-sight, slope, aspect, ridges and rivers, andvolumetrics.

TINs can model points, lines, and polygons. Atriangulation is made of many mass points, each anx,y,z tuple. Breaklines represent streams, ridges, andother linear discontinuities. Exclusion areasrepresent polygons with same elevation, such aslakes or project boundaries. Contour maps can begenerated from a TIN, using linear interpolation or asmoothing algorithm.

Implementing data representation models

A geodatabase implements the vector datarepresentation with feature datasets and featureclasses, the raster data representation with rasterdatasets, and the triangulated data representationwith triangulated irregular networks (TINs).


MODELING SURFACES

A GIS can model a surface in three general ways: asa surface raster, as contour lines, or as a triangulatedirregular network.

Each approach has merit, but the triangulatedirregular network has special analytic powers andthe surface raster can also perform interestinganalysis.

SURFACE RASTER

Some terrain data comes in the form of a uniformgrid with elevation values. An example is the DigitalElevation Model (DEM) data product from the UnitedStates Geological Survey.

6907

6899

6872

6914

6911

6892

6923

6915

6907

Surface raster Interpolated contours

6900

6920

A raster dataset can represent point elevationsspaced at regular intervals. Each cell in the raster hasan associated elevation value.

From a raster dataset with elevations, the elevationfor any point on a surface can be estimated and a setof contours can be derived.

The advantages of raster datasets are:

• The conceptual model of raster datasets is simple.Data storage is very compact.

• The raster model has well-established algorithmsto process raster data.

• Elevation data in raster format is relativelyabundant and inexpensive to obtain.

The disadvantages of raster datasets are:

• The rigid grid structure does not conform to thevariability of terrain.

• The original data is not maintained when it isinterpolated to a regularly spaced grid.

• Linear features cannot be represented well formany applications.

CONTOUR LINES

Contour lines can be used to represent surfaces. Acontour is a line following an equal elevation value.Contours are the most accessible source of terraininformation for most map users.

Contours are good for human interpretation. Closelyspaced contours are a clear visual cue that the localterrain is steep. A sharp angle in a contour is a clueof a stream or ridge line. You can get a sense of the“lay of the land” by reading contours on a map.

However, contours are generally poor for computersurface models. The collection of all points oncontours does not make a good dataset for surfaces.It is difficult to remove data artifacts introduced fromconverting contours to rasters or TINs. Convertingcontours is usually a last resort for building a surfacemodel.

You can make a perspective view or perform surfaceanalysis of contours only after they have beenconverted to a raster or a TIN.

TRIANGULATED IRREGULAR NETWORKS

A triangulated irregular network (TIN) is an efficientand accurate model for representing continuoussurfaces. TIN software includes many functions thatanalyze surfaces.

A TIN dataset is built in this way:

1. Collect a set of points with x,y,z coordinatesthrough photogrammetric instruments, GPS datacollection, or other means. Collect breaklineswhere the shape of the surface changes sharply.Collect exclusion areas for features such as lakes.


2. From this point data, GIS software creates anoptimal network of triangles, called a Delaunaytriangulation. In a TIN, each triangle is created tobe as close to equilateral as possible.

A face from a triangulated �irregular network is a triangle�floating in three-dimensional�space.��A face defines a plane, slope, �and slope direction.

z

y

x

3. Each triangle forms a face with a gradient slope.

From a TIN, an elevation can be calculated for anypoint with x and y values by first locating thetriangle and then interpolating the height inside it.

A TIN is efficient because the point density on anypart of the surface can be proportional to thevariation in terrain. A flat plain suffices with a lowpoint density and mountainous terrain requires ahigh point density, especially where the surfacechanges abruptly.

Elements of a TIN

A TIN can represent points, lines, and polygons.

�

Mass point

BreaklineExclusion �area

Project boundary

Mass points are observed spot elevations with anx,y,z coordinate triplet. They can be collected withphotogrammetric instruments, remote sensors, ordata conversion.

Breaklines delineate where the terrain has a sharpdiscontinuity in its surface. Examples of featuresmodeled as breaklines are streams, ridges, and theedges of building pads or other areas graded bymachinery.

An area of exclusion delimits an area of equalelevation. These are most commonly lakes.

Also, a project boundary can exclude the surfaceoutside an area of interest. This can be importantwhen you are calculating volumes.

Displaying surfaces with a TIN

There are several ways to visualize the surfacerepresented by a TIN. You can draw a TIN on aplanimetric (two-dimensional) map with colorsrepresenting elevation, slope, and aspect.

With three-dimensional extension software toArcInfo, you can display perspective views of asurface with draped images, contours, grid lines, orother features.

Analysis with a TIN

TIN software includes various analytic tasks on asurface. Some of the tasks are:

• Calculate the elevation, slope, and aspect (thecompass direction of slope) for any point withinthe surface.

• Generate contours by linear or quintic (smooth)interpolation across the triangulation.

• Determine a range of elevations for a surface.

• Summarize statistics for a surface, such as volumeagainst a reference plane, mean slope, area, andperimeter.

• Create vertical profile displays along alignmentson the surface.

• Perform volume calculations for roadway projects,so that the volume excavated in one area equalsthe volume deposited in another.

• Analyze which areas of a surface are visible froma point.


MODELING IMAGED OR SAMPLED DATA

Image data is collected by satellite systems or aerialphotography. Because this is by far the leastexpensive way to collect vast quantities ofgeographic data, images are an importantcomponent of many GISs.

RASTER DATASETS

Raster data can be used as a backdrop to a mapdisplay, as a source for feature extraction, forgridded surface models, or for modeling proximalgeographic functions such as dispersion. GISsoftware can rapidly overlay stacked raster datasets.

rows

columns

valueattribute

table

Value Count Land use

2

1

3

4

8

12

6

6

Agriculture

Water

Residential

Industrial

raster dataset

width

height

cell

A raster dataset stores a two-dimensional matrix withsampled values for each cell. Each cell has the samewidth and height.

The geographic coordinate of the upper-left cornerof the grid, together with the cell size and number ofgrid rows and columns, uniquely defines the spatialextent of the raster dataset.

Cell values for raster datasets can be integer orfloating numbers. Some representative types ofvalues for raster cells include:

• Light reflectance (albedo) in a photograph.

• Light intensity at a specific part of the spectrum ina satellite image.

• A derived attribute, such as land-use type, or afeature type, such as a building or street.

• A z value, such as elevation or concentration.

A value attribute table (VAT) can be optionallyassociated with a raster dataset. This table keepstrack of your value classification. You can addcustom attributes by adding more columns.

Raster datasets can have one or many bands. Eachband in a raster dataset has an identical grid layoutbut represents a different attribute. The mostcommon use of multiple bands is to represent themultispectral data captured by satellite imagery.

Raster datasets as feature attributes

Not all raster datasets have a geographic reference.An image can be used as an attribute to a feature.

If you are building a GIS to sell homes, you maywant an Internet application where the prospectivebuyer is shown a map with symbols for each homefor sale. The buyer can click a symbol to display animage, facts about the house, and the price.

Other examples of images as feature attributes are:

• Scanned documents, such as permits or deeds.

• Field data forms associated with locations.

• Blueprints or schematic diagrams of floorplans.

Representing points, lines, and polygons

point

line

area

Points in a raster dataset can be represented by oneor a few contiguous cells. Lines can be represented


by a series of cells that have a width of one or a fewcells. Polygons can be represented by a range ofcells. Although you can visually identify points, lines,and polygons in a raster dataset, it is best to do araster-to-vector conversion if you want to interactwith features.

Converting raster datasets

Raster datasets can be generated with ease, yet thefeatures they depict are sometimes more useful inanother type of dataset. An example is converting aphotograph of buildings into a feature datasetcontaining buildings with polygon shape.

The resolution of the raster dataset stronglyinfluences the accuracy of converted vectors.

Raster analysis

GIS software for raster datasets comes equipped witha powerful set of operations. These are a few:

• Spatial transformations. A raster dataset can bemoved, bent, or stretched to fit the true location.It can also be projected on a coordinate system.Rubber sheeting locally adjusts rasters to fit user-defined vectors. Polynomial transformations applyglobal equations to fit grids to user-definedvectors.

• Spatial coincidence. Modeling characteristics oflocations, such as assessing the suitability for

some type of land development like the optimallocation of a new road, or estimating land values.

distancetravel �time �

length

• Proximity. Modeling the distance to othergeographic phenomena. This distance can bemeasured as straight Euclidean distance or anabstraction, such as travel time.

301 305 315 319

317 323 328 314

305 309 312 303

• Surface analysis. Finding the qualities ofcontinuous surfaces, such as elevation, noise, orpollution concentration. You can calculate slopeand aspect from a land surface or determine thenoise level in the vicinity of an airport.

• Dispersion. Modeling the movement ofphenomena, such as simulating the spreading offire or predicting the movement of an oil spill.

• Least-cost path analysis. You can calculate theshortest path across the surface based on anydesired impedance values.


MODELING DISCRETE FEATURES

Geographic features are located at or near thesurface of the earth. Geographic features can occurnaturally (rivers, vegetation), can be constructions(roads, pipelines, buildings), and can be subdivisionsof land (counties, land parcels, political divisions).

Maps model the world with points, lines, andpolygons.

• Points represent geographic features too small tobe depicted as lines or areas.

• Lines represent geographic features too narrow tobe depicted as areas.

• Polygons represent sizeable continuousgeographic features.

An x,y (Cartesian) coordinate system references real-world locations.

FEATURE DATASETS

y

x

(3,4)

(6,11)

(7,8)

(11,3)

(15,9)

5 10 15 20 25

5

10

(19,3)

(23,9) (29,9)

(27,6)

(29,3)

In feature datasets, each location is recorded as asimple x,y coordinate. Points are recorded as asingle coordinate. Lines are recorded as a series ofordered x,y coordinates. Polygons are recorded as aseries of x,y coordinates defining line segments thatenclose an area.

Point features

Points represent geographic features that have noarea or length, or features that are too small for theirboundaries to be apparent for a given map scale.

Line features

Linear features represent objects that have length butno area, or features whose shapes are very narrow ata given map scale.

Polygon features

Polygon features are used to represent areas such asstates, counties, census tracts, sales territories, soilunits, parcels, and land-use zones.

Polygons enclose areas that meet a user-specified setof common characteristics for the phenomena beingrepresented.

How maps convey descriptive information

Maps present descriptive information aboutgeographic features using symbols and labels.

Here are some common ways that maps presentattribute information about the geographic featuresthey represent:

• Roads are drawn with various widths, patterns,and colors to represent different road classes orother attributes.

• Streams and water bodies are typically drawn inblue to indicate water.

• Special symbols denote specific features such asrailways and airports.

• City streets are labeled with names and oftenaddress ranges.

• Special buildings are labeled with their names orfunctions.


FEATURES, NETWORKS, AND TOPOLOGY

Features can have three basic roles with respect toone another: simple, networked, and topological.

Simple features

Features can be simple, with no explicit connectionsor topological associations to other features.

Simple features

pointspolylines polygonsmultipoints

Network features

Features can be connected in a network.

Geometric network

junction

edge

A network contains edges that have nodes at theirendpoints. A node can be connected to one or manyedges. This assemblage of edges and nodes is calleda geometric network.

Shared edges in a topology

The topological elements of features can be derivedand edited.

Planar topology

edgeface

node

input features

In the ArcMap editor, you can specify a set offeatures and create a planar topology, which is a setof topological primitives: nodes, edges, and faces.

When you edit a node, the connecting edges rubber-band. When you edit an edge, you are modifyingthe shape of two faces at once.

FEATURES AND CARTOGRAPHY

Features are geographic objects in the context of amap. And a map has a scale, which determines thefeature’s dimension: point, line, or polygon.

Buildings can be drawn as polygons at large scaleand points at small scale.

small scale

large scale

* *

**

** ***

*** * **

***

A stand of trees might be drawn individually at largescale, but a forest is drawn as polygons that boundtrees above a certain density.

large scale small scale

A stream system at large scale will have many shapepoints and minor branches. At small scale, the linedetail is filtered and minor streams are removed.

If you need to change the feature dimension atvarying scales, you can set up a database relate toanother feature class. In this case, trees areassociated with a forest stand. When you draw amap, the scale determines which set of features isdrawn.


COMPARING SPATIAL DATA REPRESENTATIONS

Sources ofdata

Vector datarepresentation

Raster datarepresentation

Triangulated datarepresentation

Compiled from aerial photographyCollected from GPS receiversDigitized from map manuscriptsSketched on top of raster displayVectorized from raster dataContours from triangulationReduced from survey field dataImported from CAD drawings

Photographed from an airplaneImaged from a satelliteConverted from a triangulationRasterized from vector dataScanned blueprints, photographs

Compiled from aerial photographyCollected from GPS receiversImported points with elevationsConverted from vector contours

Spatialstorage

Featurerepresentation

Topologicalassociations

Geographicanalysis

Points stored as x,y coordinates.Lines stored as paths of connectedx,y coordinates. Polygons stored asclosed paths.

From a coordinate in the lower-leftcorner of the raster and cell heightand width, each cell is located by itsrow and column position.

Each node in a triangle face has anx,y coordinate value.

Topological map overlayBuffer generation and proximityPolygon dissolve and overlaySpatial and logical queryAddress geocodingNetwork analysis

Spatial coincidenceProximitySurface analysisDispersionLeast-cost path

Elevation, slope, aspect calculationsContour derivation from surfaceVolume calculationsVertical profiles on alignmentsViewshed analysis

Line topology keeps track of whichlines are connected to a node.Polygon topology keeps track ofwhich polygons are to the right andleft sides of a line.

Neighboring cells can be quicklylocated by incrementing anddecrementing row and column values.

Each triangle is associated with itsneighboring triangles.

Points represent small features. Linesrepresent features with a length butsmall width. Polygons representfeatures that span an area.

Point features are represented by asingle cell. Line features arerepresented by a series of adjacentcells with common value. Polygonfeatures are represented by a regionof cells with common value.

Point z values determine the shape ofa surface. Breaklines define changesin the surface such as ridges orstreams. Areas of exclusion definepolygons with the same elevation.

Cartographicoutput

Vector data is best for drawing theprecise shape and position offeatures. It is not well suited forcontinuous phenomena or featureswith indistinct boundaries.

Raster data is best for presentingimages and continuous features withgradually varying attributes. It is notgenerally well suited for drawing pointand line features.

Triangulated data is best for richpresentation of surfaces. This datacan be viewed by using color to showelevation, slope, or aspect or in athree-dimensional perspective.

Focus ofmodel

Vector data is focused on modelingdiscrete features with precise shapesand boundaries.

Raster data is focused on modelingcontinuous phenomena and imagesof the earth.

Triangulated data is focused on anefficient representation of a surfacethat can represent elevation or otherquality, such as concentration.


In summary, there are three basic representations ofspatial data: vector, raster, and triangulated. Each ofthese representations has merits and is well suitedfor a particular class of geographic analysis andcartographic output.

These spatial data representations are not exclusive;your geodatabases can contain all three for the mapuses for which they are best suited. A map can displayany or all of these spatial data representations.

Often, raster data is displayed as a background layerto vector data. This provides a photo-realistic contextto the vector layers on which you might be perform-ing engineering or analysis.

Triangulated data is also sometimes drawn as abackground layer to vector data to provide avisualization of the shape of the earth’s surface.

CHOOSING A SPATIAL DATA REPRESENTATION

You must consider many issues when choosing aspatial data representation. Often, the choice is clearand guided by the available data and the analyticaltasks you need to perform. Sometimes, it is not soobvious which data representation is best.

Surfaces are a good example; two robust ways torepresent a surface are as raster data or astriangulated data. A choice requires more study.

The following are a few considerations for choosinga spatial data representation.

Is the focus on features or location?

If you are modeling distinct objects with attributesand behavior, the vector data representation issuperior.

If you are modeling continuous objects orphenomena characterized by an attribute at alocation, you should choose between raster ortriangulated data.

Raster data models an area with uniform sampling ofattributes in a regular grid. Triangulated data modelsan area with points and values sampled at a variabledensity.

What data is readily available?

A major influence on your selection of a datarepresentation is what data is already available.

An early step in the design of your GIS is a survey ofall the geographic data already available. When youfind the data that is most suitable, you will make ajudgment on whether that data is sufficient orwhether you will need to create new data by othermeans such as aerial photography, GPS datacollection, or digitization.

Sometimes, you might choose to convert existingdata from one representation to another. Forexample, the best source for electric transmissionlines might be scanned maps in raster format. Toperform electrical analysis or environmental studies,you may find it necessary to convert that raster datainto vector data. You will weigh the cost and qualityof output of this raster-to-vector conversion byanother means of data collection.

What is the required precision for locatingfeatures?

If you need to locate features with significantprecision, you should choose vector datarepresentation. Feature identification and selection iseasier with vector data, and precise coordinatevalues are stored.

Determining locations of features in raster data isconstrained by the dimensions of each cell. Intriangulated data, only the locations of points andbreaklines are well defined. The locations of featuresand their shapes are generally indistinct in raster andtriangulated data.

What types of features are required?

If you are modeling large features with values thatvary, change with time, or have indistinctboundaries, the raster data representation is usuallybest. An example is the modeling of a fire over timeor the dispersion of groundwater contaminants.

If you are modeling features that characterize theshape of the earth’s surface, such as mountain peaks,ridge lines, or stream lines, the triangulated datarepresentation is often best.


Some natural features are better represented withvector data. An example is a river system. If you aredisplaying rivers as a background layer on a map ormodeling ship traffic on a river as part of a broadertransportation analysis, you will probably choose thevector data representation.

If you are modeling man-made features, the vectordata representation is most often best. Man-madefeatures have well-defined shapes that arecharacterized by straight lines and circular arcs. Also,man-made features are often located with survey-level precision.

What type of topological association is desired?

Some objects are nontopological and can be freelyplaced in a geographic area. For example, an areadefining a wildlife habitat is arbitrary, ill-defined, andoverlaps other habitats, and so it does not have atopological relationship with other features.

Also, many objects are primarily stored in a GIS forthe purpose of background display on a map, so it isusually not necessary to store them in a topologicalformat. If roads are a background layer in your GIS,they will probably be simple features. If roads arepart of an analysis of a transportation system, theyshould be topological features.

A GIS can have networks and topologies and theseare captured within the vector data representation.Networks represent roads, rivers, and utilities.Topologies represent collections of areas where eachpoint in an area is covered by exactly one polygon.

What type of analysis is required?

If you are analyzing a surface, the triangulated datarepresentation supports the broadest array of analyticfunctions. However, the raster data representationalso represents some surface-modeling functions.

The triangulated data representation supportsvolume calculations between two surfacesrepresenting undeveloped and developed areas,what area is visible from a point in space, thedetermination of elevation, slope, and aspect for anypoint on a surface, and the generation of verticalprofiles of an alignment, such as road or utility.

If you are analyzing dispersion of an indeterminatefeature over time, such as a plume of pollution, thenyou should select the raster data representation. Theraster data representation also supports thedetermination of proximity to features, least-costpaths, and rapid overlay of rasters for suitabilityanalysis.

If you are locating optimal locations for placing abusiness or performing a service, studying flowsthrough a network, managing land records,referencing postal addresses to a location on a map,or querying features on a map, you should choosethe vector data representation.

The vector data representation allows analysis that isbased on spatial relationships such as proximity andadjacency, and topological relationships such asupstream and connected.

What types of maps are to be produced?

The type and quality of the desired cartographicpresentation also guides which spatial datarepresentation is recommended.

The raster and triangulation data representationsproduce attractive maps of areas with varyingattribute values. The vector data representationmakes maps with fine detail for features.

Cartographic considerations will further guidewhether points, lines, or polygons for vector datarepresentation are best. For example, the map scalewill guide whether buildings should be representedas points or polygons, or rivers as lines or polygons.

CONCLUSION

ArcInfo provides a rich infrastructure for the threefundamental representations of geography. The nextchapter reveals how geographic data is structuredand presented in the ArcInfo applications.

61

4Chapter intro text

• Chapter intro - bullet list (use 12 pt bulletsize)

< Align text with this guide.

The structureof geographicdata

Sets of geographic data are organized in yourcomputer’s file system and databasemanagement system. The catalog synthesizesthese two sides of data organization andpresents a unified user interface and datamodel. The catalog also makes it easy to workwith local and networked data.

In this chapter:

• The catalog and connections to data

• The geodatabase, datasets, and featureclasses

• ArcInfo workspaces and coverages

• Shapefiles and CAD files

• Maps and layers

• Comparing the structure of vectordatasets

• Comparing feature geometry in vectordatasets

Ionian Isles and Greece, John Rapkin, 1851.


THE CATALOG AND CONNECTIONS TO DATA

Your desktop computer has all sorts of dataorganized into folders, documents, spreadsheets,and databases. You use documents for letters andreports, spreadsheets for expense reports, anddatabases to keep inventories of customers andproducts. And you organize files into a folderhierarchy that makes sense; you might organizethem by clients, projects, time span, or anymeaningful association.

Likewise, a geographic information system (GIS)manages data in a hierarchy of folders, files, andgeodatabases. The primary types of geographic data—vector, raster, TIN, locations—can be containedwithin geodatabases or files.

Your geographic data can be hosted in a single-usergeodatabase on your computer’s disk or a multiusergeodatabase hosted on a database server. And youcan structure your geodatabases and folders toreflect project areas, thematic groupings, departmentorganization, or other ordering.

THE CATALOG

ArcCatalog is the ArcInfo application that lets youexplore, access, manage, and build your geographicdata. It presents your geographic data in a mannersimilar to Microsoft’s Windows Explorer.

The items you see in the catalog represent dataobjects such as geodatabases and feature classes,map objects such as maps and layers, and ancillaryobjects such as styles and coordinate systems.

The collection of connections to geographic data iscalled a catalog. A catalog provides you with aseamless view of geographic data—file-based dataand personal databases are located within arecognizable tree hierarchy. A catalog can also drillinto relational databases and reveal some of theirinternal structure, particularly the tables that storegeographic data.

The catalog reveals the structure of geographic datathrough special icons that communicate the role ofthe various elements of your geographic database.Some catalog items represent folders and files inyour Windows file system. Other items representcollections of features and objects withingeodatabases. Certain items are references to

geodatabases and relational databases accessedacross a network.

Some tasks you can perform with a catalog include:

• Creating and formatting new data

• Searching for data

• Assessing geographic extent and suitability of data

• Documenting the provenance and quality of data

• Launching GIS operations

• Publishing data for widespread access

Single-user and multiuser geodatabases

Geodatabases come in two basic variants—personalgeodatabases using Microsoft Access .mdb files andmultiuser geodatabases served through ESRI’s ArcSpatial Database Engine (ArcSDE) to one of theleading relational database management systems.

These two variants are functionally identical exceptthat multiuser geodatabases support versioning,which allows multiple users to access and edit thecommon geographic database. Versioning isdiscussed further in chapter 7, “Managing WorkFlow with Versions.”

Folder connections and database connections

Folder connections and database connections giveyou a consistent and unified view of all your data.

A folder connection lets you access data on yourlocal drives or shared drives on networkedcomputers.

A database connection contains the specificationsfor accessing a database: server or IP address,instance or TCP port information, and account username and password. You can access geographicdata in a relational database management systemthrough ArcSDE or you can access nonspatialattribute data through an ODBC (Open DatabaseConnectivity) driver.

Once you connect to a remote multiusergeodatabase and expand its tree nodes in thecatalog, you will see exactly the same structure ofconstituent data objects on folder connections.

Chapter 4 • The structure of geographic data • 63

A folder connection points to a root folder on a local disk or a selected folder on a diskaccessed through a network.

All folders under each folder connection appear in the catalog. If the folder does notcontain geographic datasets, it appears as an ordinary folder.

The catalog is the presentation of maps and geographic data available on your local and network disks.It can be customized to reference special locations of data and supporting file types.

Folders that contain geographic data are shown with a special icon. They contain geodatabases, coverages, shapefiles, CAD files, and related files.

Database connection folders contain connections to relational databases and geographic datahosted in multiuser relational databases.

A geodatabase is an object-oriented store of geographic data. When located inside a folder, it is a personal geodatabase.

A coordinate systems folder contains a number of coordinate system files.

A coordinate system file contains the mathematical specification for how an area istransformed from a spheroid to a planar coordinate system.

A database connection is a reference to a geodatabase stored in a multiuser database management system and accessed through ArcSDE or a nongeographicrelational database accessed through ODBC.

A database connection wizard walks you through the steps to establish and test adatabase connection to an ArcSDE or OLE DB database.

The catalog, folders, andconnections


THE GEODATABASE,DATASETS,AND FEATURE CLASSES

When you design and implement your geographicdata model, you have considerable discretion todesign each level of your file management systemand database schema. The catalog can adapt toyour existing organization of data or you can devisea new structure optimized for access andadministration.

ORGANIZING GEOGRAPHIC DATA

Geographic data is organized into a hierarchy ofdata objects. In your data design process, you canorganize your data by work groups, thematic type,common spatial extent and coordinate system, ortopological associations.

Geodatabases

A geodatabase is the top-level unit of geographicdata. It is a collection of datasets, feature classes,object classes, and relationship classes.

Your sum aggregate of geographic data can spanone, several, or many geodatabases. Geodatabasesare usually organized into broad categories of datasuch as land base, transportation, environment, andutility infrastructure.

Geodatabases manage seamless geographic data.There is no partitioning of a geographic area intotiled units. Rather, geodatabases use effective spatialindexing for continuous representation of an extent.

Personal geodatabases can represent small- tomedium-sized datasets. Very large datasets can beefficiently handled with an enterprise ArcSDEimplementation.

Geographic datasets

There are three general types of geographic datamodels: vector, raster, and triangulation. In thegeodatabase, they are implemented by three typesof geographic datasets: the feature dataset, the rasterdataset, and the TIN dataset.

A feature dataset is a collection of feature classesthat share a common coordinate system. You maychoose to organize simple feature classes inside oroutside of feature datasets, but topological featureclasses must be contained within a feature datasetto ensure a common coordinate system.

A raster dataset can either be a simple dataset or acompound dataset with multiple bands for distinctspectral or categorical values.

A TIN dataset contains a set of triangles that exactlyspan an area with a z value for each node thatrepresents some type of surface.

Object classes

An object class is a table in a geodatabase withwhich you can associate behavior. Object classeskeep descriptive information about objects that arerelated to geographic features, but are not featureson a map.

An example of an object class is owners of landparcels. You can establish a database join betweena polygon feature class for land parcels and anobject class for owners.

Feature classes and topology

A feature class is a collection of features with thesame type of geometry: point, line, or polygon. Youcan think of two categories of feature classes—simple and topological.

Simple feature classes contain points, lines,polygons, or annotation without any topologicalassociations among them. That is, points in onefeature class may be coincident with, but distinctfrom, the endpoints of lines in another featureclass. These features can be edited independentlyof each other.

Topological feature classes are bound within agraph, which is an object that binds a set of featureclasses that comprise an integrated topological unit.ArcInfo 8 introduces the first type of graph in ageodatabase—geometric networks.

Relationship classes

A relationship class is a table that storesrelationships between features or objects in twofeature classes or tables. Relationships modeldependencies between objects.

With relationships, you can control what happensto an object when its related object is removed orchanged.


A feature dataset is a collection of feature classes, graphs, andrelationship classes that share a common spatial reference.

A line feature class is a collection of simple features with polylinegeometries.

A polygon feature class is a collection of simple features withpolygon geometries.

An object class is a table with behavior. It is a matrix of rows that represent objects and columns that represent attributes.

A junction feature class contains simple or complex junctionfeatures that participate in a geometric network.

An edge feature class contains simple or complex edgefeatures that participate in a geometric network.

A raster dataset represents imaged or sampled data on a rectangulargrid. It can have one or many raster bands.

A geometric network defines a set of junction and edgefeature classes that collectively form a one-dimensionalnetwork.

A point feature class is a collection of simple features with pointor multipoint geometries.

A folder connection.

A folder with geographic data.

A geodatabase is a store of geographic data organized into geographic datasets and feature classes. A geodatabase under a folder is a single-user geodatabase.

A relationship class is a collection of relationships betweenfeatures in two feature classes.

Feature classes withsimple geometry

types and tables canbe placed directly

under ageodatabase orunder a feature

dataset.

Geometricnetworks and

network featureclasses must be ina feature dataset.

A catalog.

A database connection folder lets you access multiuser geodatabases served byArcSDE.

When you expand a database connection that represents a multiuser geodatabase, itcontains the same types of datasets and feature classes as a single-user geodatabase.

Relationship classes can beplaced in a feature dataset

or directly in ageodatabase.

Behavior for featureclasses and object

classes isimplemented bydefining rules orextending a classand writing code.

The catalog’s view of ageodatabase


ARCINFO WORKSPACES AND COVERAGES

For many years, ArcInfo coverages have been usedto represent vector data. The coverage format hasenjoyed widespread implementation atgovernmental agencies, corporations, andorganizations throughout the world because itefficiently stores spatial and topological data;attribute data is stored in relational tables that canbe customized and joined with other databases.

Coverages combine spatial data and attribute dataand store topological associations among features.Spatial data is held in binary files and attribute andtopological data is kept in INFO™ tables. Thecatalog combines the representation of coveragebinary files and INFO tables into coverage featureclasses.

The introduction of geodatabases in ArcInfosupplements but does not replace coverages. Wherethey are employed, coverages still most often servethe intended requirements of applications.Coverages can be displayed, queried, analyzed, andedited in the new ArcInfo applications.

You can choose to migrate coverages intogeodatabases when the benefits of integratingfeature behavior and storing all data in a databaseoutweigh the effort of conversion. You can think ofa geodatabase as a next-generation coverage.

WORKSPACES AND GEOGRAPHIC DATA

ArcInfo workspaces contain the three basicrepresentations of geographic data—coveragescontain vector data, grids contain raster data, andTINs contain triangulations that represent surfaces.Most data stored in a workspace implements thegeorelational model where topology is stored andattributes are linked to features.

An ArcInfo workspace is a special type of folderwhere attributes for data are stored in INFO tablesand all of the tables are managed through an INFOsubfolder that is invisible in the catalog. When youuse the catalog to create, move, and delete items inan ArcInfo workspace, their integrity is maintainedfor you. You should never use Windows Exploreror My Computer to manage coverages, grids, orTINs; the synchronicity between coverages and theINFO subfolder will be broken and data corrupted.

COVERAGES, FEATURES, AND TOPOLOGY

Coverages contain feature classes that arehomogeneous collections of features.

Tic

Arc Text

Polygon

Node

Labelpoint

Originalfeatures

Rubber-sheetedfeatures

Links

A region is a compositefeature of polygons.

Region of disjointpolygons

Region oftouchingpolygons

Region of nestedpolygons

A route is a composite feature ofsections. A section is a whole or

partial arc.

Routes can be simplyconnected, disjoint,looped, or intersecting.

The primary types of coverage features are points,arcs (lines), polygons, and nodes. These featureshave topological associations: arcs form theperimeter of polygons, nodes form the endpoints ofarcs, points mark the interiors of polygons. Pointfeatures have a dual identity; they can representsmall geographic objects such as wells andbuildings and they can mark polygon interiors.

Secondary types of coverage features are tics, links,and annotation. Tics are used for map registration,links are used for adjusting features, and annotationis used to label features on a map.

Coverages also contain composite features. Routesare collections of arcs with an associatedmeasurement system. A common use of routes isfor transportation systems. Regions are collectionsof polygons that can be adjacent, disjoint, oroverlapping. Regions are used for land-use andenvironmental applications.


A coverage link feature class contains vectors used for adjusting localareas to known reference points. This is called rubber sheeting.

A coverage label feature class contains label points that mark polygons.Each polygon contains one label point.

A coverage node feature class contains node features that occupy theends of arcs. Attributes are kept in node attribute tables (NAT) .

A coverage tic feature class contains points used for registering maps thatare digitized.

A coverage annotation feature class contains text on a map. Attributescan optionally be kept in text subclass tables.

A point coveragecontains a pointfeature class. It canoptionally containtic, link, andannotation featureclasses.

A coverage point feature class contains point features. Attributesare kept in point attribute tables (PAT).

A coverage arc feature class contains line features that form a network ordefine polygon boundaries. Attributes are kept in arc attribute tables (AAT) .

A line coveragecontains an arcfeature class. It canoptionally containnode, route, point,tic, link, andannotation featureclasses.

A coverage region feature class contains composite areal features formedby a number of polygons. Attributes are kept in region subclass tables .


A folder with coverages, grids, and TINs is called an ArcInfo workspace . Attributesfor most of these feature classes are stored in INFO tables.

A coverage is an integrated set of feature classes that contain topology. Thisicon denotes a coverage with polygon topology.

A TIN is composed of points with z-values organized into a mosaicked setof triangles to represent a surface.

A grid is a rectangular matrix of cells that represent imaged or sampleddata. Attributes for values are kept in a value attribute table (VAT).

An INFO table is a relational database table. Feature attribute tables areINFO tables linked with features by object identifiers.

A catalog.

The catalog’s view of anArcInfo workspace

A coverage route feature class contains composite line features with alinear measurement system. Attributes are kept in route subclass tables.

A coverage polygon feature class contains areal features formed by a ring of arcsand interior label points. Attributes are kept in polygon attribute tables (PAT) .

A polygon coveragecontains a polygonand label pointfeature class. It canoptionally containregion, arc, node,route, tic, link, andannotation featureclasses.


SHAPEFILES AND CAD FILES

While topological datasets such as geodatabasesand coverages provide a foundation for richgeographic analysis and map display, many mapuses can be satisfied with a simpler form of featuredata.

Simple feature classes store the shapes of featureswith points, lines, and polygons, but do not storetopological associations. This structure has theadvantage of simplicity and rapid displayperformance, but the disadvantage of not beingable to enforce spatial constraints.

For example, if you are making a parcel map, youwant to ensure that the polygons forming parcelsdo not overlap or have gaps between them. Simplefeature classes cannot ensure this type of spatialintegrity.

Yet, simple feature classes comprise a large part ofavailable geographic data because they are easy tocreate and are sufficient for geographic data thatforms background layers on maps.

The geodatabase can contain simple feature classes.ArcInfo also supports interaction with shapefiles andCAD drawings in AutoCAD® and MicroStation®

format, common repositories of simple feature data.

SHAPEFILES

ArcView GIS 2, an ESRI software product for mapdisplay and query, introduced the shapefile formatto satisfy the need for simple feature datasets.

A shapefile is composed of three main files thatcontain spatial and attribute data. A shapefile canoptionally have other files with index information.In the catalog, all these files that comprise ashapefile appear as one feature class.

Simple polylinePoint

Multipoint Multipart polylines

Simple polygon

Multipart polygons

A shapefile is a homogeneous collection of featuresthat can have either point, multipoint, polyline, orpolygon shapes.

A point shapefile contains features with pointgeometry. A point is a single coordinate value.

A multipoint shapefile contains features withmultipoint geometries, in which several pointsrepresent one feature.

A line shapefile contains features with polylinegeometry. Polylines are made of paths, which aresimply connected sets of line segments. The pathsin a polyline can be connected, disjoint, orintersecting.

A polygon shapefile contains features with polygongeometry. A polygon contains one or many rings. Aring is a closed path that cannot intersect itself. Therings in a polygon can be disjoint, nested, orintersect one another.

While shapefiles store attributes in an embeddeddBASE file, attributes of other objects can be storedin another dBASE table and can be joined to ashapefile by an attribute key.

CAD DRAWINGS

A substantial amount of geographic data has beencollected in CAD (computer-aided design) drawingfiles. A characteristic of CAD files is that features aretypically subdivided into many layers.

“Layer” in a CAD file has a different meaning than“layer” in a map. In a CAD file, it represents a set ofsimilar features. In a map, it represents a referenceto a geographic dataset or feature class with anassociated drawing method.

A CAD dataset is the catalog’s representation ofCAD drawing files. It is subdivided into CAD featureclasses, each of which aggregates all of the layersfor points, lines, polygons, or annotation. If a CADdataset has 17 layers—three with points, eight withlines, four with polygons, and two with annotation,they will be combined into a CAD point featureclass, a CAD line feature class, a CAD polygonfeature class, and a CAD annotation feature class.

ArcInfo supports interaction with CAD files incertain AutoCAD and MicroStation formats. Consultthe online help for details on supported CADformats.


A line shapefile.

A polygon shapefile.

A dBASE table is a database table with rows and columns that can bejoined to a shapefile. (Shapefiles contain embedded dBASE tables.)


A folder with shapefiles and CAD files.

A point or multipointshapefile.

A catalog.

A CAD drawing is a representation of an entire CAD dataset that can beplaced on a map, but not drawn as a layer.

A CAD dataset contains point, line, polygon, and annotation layers. ACAD dataset can be in AutoCAD or MicroStation format.

A CAD point feature class is the set of CAD layers in a CADdataset that contains points.

A CAD line feature class is the set of CAD layers in a CAD datasetthat contains lines.

A CAD polygon feature class is the set of CAD layers in a CADdataset that contains polygons.

A CAD annotation feature class is the set of CAD layers in a CADdataset that contains text.

Shapefiles are collections of point,multipoint, line, or polygon features.Shapefiles contain simple features; thereare no topological associations amongthese features. Attributes for shapefilesare kept in dBASE tables.

The catalog’s view of shapefiles andCAD drawings


MAPS AND LAYERS

The catalog not only provides access to geographicdata, it lets you manage the files that store yourmap and layer definitions. These files permanentlystore all the cartographic specifications you make inyour maps.

Maps and layers in the catalog make it possible foryou to create high-quality maps without writing anymacro code. They also enable your organization tostandardize the format, content, and appearance ofyour finished maps.

MAP DOCUMENTS, TEMPLATES, AND STYLES

Whenever you create a map in ArcMap, it is storedas a file on your computer disk with a .mxd fileextension. This is called a map document.

A map stores the cartographic elements thatcomprise a map, but does not store geographicdata. Instead, the layers in the map referencegeographic datasets located at any location on adisk or database accessible by the catalog.

A map document template is a starting point formaking any kind of map. It can be fairly simplewith a set page size and style or it can be morecomplex with many cartographic elements andlayers predefined. Templates make it easy for youto repetitively create a series of maps with aconsistent appearance.

A style is a collection of palettes of cartographicobjects that you use to draw maps. These objectsinclude the marker symbols you use to draw pointfeatures, line symbols to draw line features, fillsymbols to draw polygon features, and text symbolsto draw annotation. Other objects in a style includecolors and certain cartographic elements such asnorth arrows.

The purpose of styles is to ensure consistent useof cartographic symbols on your maps. Yourorganization can have multiple styles for generatingdifferent types of map products.

LAYERS

Layers can either be stored within a map documentor in a separate layer file with a .lyr file extension.When you make simple maps, it will be mostexpedient for you to simply create your layers

within the map, but when you want to share layerswith other people, it is better to create them asseparate layer files.

Since layers are references to geographic data,when a geographic dataset is moved or renamed inthe catalog, simply updating the layer with the newlocation of the geographic data guarantees that allmaps that include that layer still have the correctreference to data.

Map and layer files can be stored anywhere onyour computer or network. They can be organizedin a folder with geographic data or in a separatefolder of their own.

The layers that you see in the catalog are only thelayers that are stored in stand-alone files. To see thelayers embedded in a map document, you will goto the table of contents for a map in the ArcMapapplication.

Vector, raster, and TIN layers

Point, line, and polygon layers can reference anyfeature class with zero-, one-, or two-dimensionalfeature geometries.

Point layers can reference points and junctions in ageodatabase; label points, tics, and nodes in acoverage; and points in a shapefile or CAD dataset.

Line layers can reference lines and edges in ageodatabase; arcs and routes in a coverage; andlines in a shapefile or CAD dataset.

Polygon layers can reference polygons in ageodatabase; polygons and regions in a coverage;and polygons in a shapefile or CAD dataset.

Annotation layers can reference text in ageodatabase or coverage.

Raster layers can be referenced by grids in anArcInfo workspace, and image files in a variety offormats.

TIN layers can be referenced by TINs in an ArcInfoworkspace.



A folder with maps and layers.

A map document stores the cartographic elements that make up a map:title, north arrow, legend, scale bar, data frames, and layers.

The catalog.

A map document template contains preset cartographic elements toimplement mapping standards. Map documents are made from templates.

A group layer is a collection of layers that are grouped for convenientplacement on a map.

A point layer references a feature class with single-coordinate featuresfrom geodatabases, coverages, shapefiles, CAD files, or other datasets.

A line layer references a one-dimensional feature class such as a line or edgefeature class in a geodatabase, a coverage arc feature class, or a line shapefile.

A polygon layer references a two-dimensional feature class such as apolygon feature class in a geodatabase or coverage or a polygon shapefile.

An annotation layer references an annotation feature class in ageodatabase, coverage, or CAD dataset.

A TIN layer references a TIN dataset in a geodatabase or ArcInfo workspace.

A raster layer references a raster in a geodatabase, a grid in an ArcInfo workspace,or an image file in a folder.

A CAD layer references a CAD drawing in several common formats.

The catalog’s view ofmaps and layers

Layers can either beresident within a

map document orcan be a stand-alone

file in any folder.


COMPARING THE STRUCTURE OF VECTOR DATASETS

Datacollection

Featuredataset

Featureclass

A geodatabase is a collection of �feature datasets, rasters, and TINs.

All spatial, topological, and attributedata is stored in tables in a relationaldatabase.

Geodatabases span continuousgeographic extents.

Behavior is tightly coupled withfeatures through rules and codewritten for custom feature classes.

The three major types of geographic datasets that you work with in ArcInfo are geodatabases, coverages, andshapefiles. Because of how they are implemented in folders and databases and how topological information isstored, their data objects are structured differently.

This table summarizes the structure of vector data objects in geodatabases, coverages, and shapefiles and thefunctions they support. If you already use coverages and shapefiles, this table clarifies how they compare withgeodatabases.

Geodatabase Coverage Shapefile

A feature class stores features in a�relational table with a special field forthe geometric shape of a feature.

The types of feature classes are point,line, polygon, annotation, simplejunction, complex junction, simple edge, and complex edge.

A feature class can be extended to acustom feature class.

A coverage feature class stores feature geometry in a binary file andattributes and topology in a featureattribute table.

The primary coverage feature classesare point, arc, polygon, and node.Secondary feature classes are tic,link, and annotation. Compoundfeature classes are region and route.

A coverage feature class cannot beextended.

A feature dataset in a geodatabase �contains simple or topological featureclasses.

Line topology is implemented througha geometric network. Polygon�topology is implemented through on-the-fly topological editing.

Many feature classes can beassociated with a topological role.

User-defined associations can beestablished between features indifferent feature classes.

Feature datasets have a definedcoordinate system.

A coverage contains topological �feature classes that participate in lineor polygon topology.

Line topology is implemented witharcs, nodes, and routes. Polygontopology is implemented with arcs,label points, polygons, and regions.

Only one feature class is associatedwith a topological role.

No associations are defined except fortopological associations among relatedfeatures like arcs and polygons.

Coverages have a defined coordinatesystem.

A shapefile has one simple feature�class.

Polygon topology among a set ofshapefiles can be implemented withon-the-fly topological editing.

There is no implicit topological role fora shapefile.

No associations are establishedamong features in shapefiles.

Shapefiles have no definedcoordinate system.

An ArcInfo workspace is a collection of�coverages, grids, and TINs.

Spatial data is stored in binary files.Topological and attribute data isstored in INFO tables.

For large datasets, coverages aresubdivided into tiles in a map library.

Behavior is loosely coupled withfeatures through AML scripts or VBAmacros.

A folder can contain shapefiles. �

Spatial data is stored in binary files.Attribute data is stored in dBASEtables. No topological data is stored.

Shapefiles are continuous for small tomoderately sized datasets.

Behavior is loosely coupled withfeatures through VBA macros.

A shapefile stores feature geometries �in a binary file and attributes in adBASE file.

The types of shapefiles aremultipoint, point, line, and polygon.

A shapefile cannot be extended.


COMPARING FEATURE GEOMETRY IN VECTOR DATASETS

Pointfeatures

Linefeatures

Polygonfeatures

A feature class can containfeatures with point shapes or

multipoint shapes. A multipoint isa set of points that represent one

feature.

Network junction features are alsopoints.

Point Multipoint Label points are pointfeatures or centroids of polygonswith attributes. In a coverage withpolygon topology, each polygonmust have exactly one label point.

Nodes are endpoints ofarcs. They can have attributes.

Tics are for registration.

A coverage cannot containmultipoint features.

A shapefile can havefeatures that are simple

points or multipoints.

Points have no associationwith polygons.

PointMultipoint

A shapefile has polylinesthat have one or many

paths.

There are no topologicalassociations in a shapefile.

Polyline withone path

Polyline withseveral paths

line

circular arc

elliptical arc

Bézier curve

In a geodatabase,a polyline has oneor many paths.

A geometric network containsjunctions and edges that form a

one-dimensional network.

Paths arecomposed offour types ofsegments:

A polygon featureclass is a planargraph with simplepolygons.

A region subclassis a composite ofpolygon features.

A planar graph is a continuousmap of an area by nonintersectingpolygons. Each point in an area iscovered by exactly one polygon.

Each polygon has a label point,often at the centroid. Attributesare associated with label points.

ArcInfo has a geometry model that represents the shapes of features in geodatabases, coverages, andshapefiles. The principal building blocks of this model are points; segments that can be straight,

circular, elliptical, or Bézier curves; paths that are a set of connected segments; and rings that are aclosed nonintersecting path. A polyline is composed of one or many paths. A polygon is composed of

one or many rings.

Features in a geodatabase implement the full geometry model. Features in coverages and shapefilesimplement a subset of the geometry model.

Geodatabase Coverage Shapefile

Arcs are simply connected sets ofstraight line segments with nodes at

the endpoints.

Arcs also participate in 2-D topology.They carry information about whichpolygons are to the right and left.

Routes are composites of manysections. A section is a whole orpartial arc. Routes have arbitrary

connectivity.

L R

Polygon withnested rings

Polygon withone ring

Polygon withdisjoint rings

A polygon is made from one ormany rings. A ring is a closed,

nonintersecting path. Likepolylines, polygons can have

lines, circular arcs, elliptical arcs,and Bézier curves.

Polygon withnested rings

Polygon withone ring

Polygon withdisjoint rings

Polygons in shapefiles arestructurally similar to polygonsin geodatabases except thatthe segments can only be

straight lines.

75

5 Smartfeatures

Objects in the world have natural rules andrelationships that they follow. Rivers flowdownstream, roads handle a level of traffic,and parcels of land respect covenants. WithArcInfo, you can express this behavior throughvalidation rules on feature subtypes. Theseare the topics in this chapter:

• The qualities of features

• Steps to making features smart

• Designing the geodatabase

• Storing data in tables

• The shape and extent of features

• Attributes: qualities of an object

• Adding simple behavior with subtypes

• Validating attributes

• Relationships among objects

• Extending object classes

• The geodatabase object modelMünchen, G. Bodenehr, date unknown.


Geographic objects in the world exist within a richcontext. They occupy a position, delineation, orarea; have neighboring objects; may influencesurrounding objects as a consumer or provider of aresource; have attributes that can be values, counts,categories, or descriptions; and might have apredictable response to an external stimulus.

FEATURES IN THE GEODATABASE DATA MODEL

Features, as they are represented in the geodatabasedata model, have many qualities such as vectorshapes, relationships, attributes, and behaviors.These qualities collectively express the rich contextthat geographic objects experience.

For many applications, vector features are the mostversatile geographic data representation, suited forgeographic objects that have distinct boundaries andare persistent. Other geographic objects that can beconsidered continuous phenomena are bettermodeled with rasters or TINs.

This chapter discusses the qualities that make vectorfeatures smart in the geodatabase. First, a briefoverview.

Features have shapes

The shape of a feature is stored as a special field ina feature class table of type geometry. A feature canbe represented by one of these types of geometries:

• Points and multipoints, which are a set of points.

• Polylines, a set of line segments that may or maynot be connected.

• Polygons, a set of rings that can be disjoint orembedded. A ring is a set of connected, closed,nonintersecting line segments.

The line segments that make up polylines andpolygons can be straight line segments, circulararcs, Bézier curves, and elliptical arcs.

x-coordinate axis

y-coordinate axis

THE QUALITIES OF FEATURES

Features have a spatial reference

The shape of a feature is stored with x and y valuesin a Cartesian coordinate system. But the surface ofthe earth is roughly spherical. A spatial referencespecifies how the x,y coordinates of a set offeatures are mapped onto the earth’s surface.

Features have attributes

A feature maintains its attributes as fields in afeature class table. Feature class tables are tables ina relational database. Attributes define standard andcustom properties of features and can be numeric,textual, or descriptive.

816 High Street 1888 2200 4 bedrooms

Address BuiltSquarefootage

Number ofbedrooms

Features have subtypes

Features are collected into feature classes. Featureclasses are homogeneous sets of features, but theremay be considerable variation among features.

A feature class comprising buildings can be logicallysubdivided into subtypes such as residential,commercial, and industrial. Subtypes give youincreased control of other qualities of features suchas attribute domains and rules.

Chapter 5 • Smart features • 77

Features have relationships

All geographic objects have some relationship toother objects. You can define explicit relationshipsamong geographic objects in different featureclasses.

You can also define relationships to nonspatialobjects, such as the relationship between a houseand its owner.

Feature attributes can be constrained

To enhance the accuracy of data collection, eachattribute of a feature can have an attribute domain,which is a numeric range or a list of valid values.Each attribute can also have a default valueautomatically assigned when a feature is created.You can set distinct attribute domains and defaultvalues for each subtype in a feature class.

2200 4 bedrooms

Squarefootage

Number ofbedrooms

The square footage can beconstrained to a numericrange of between 100 and100,000.

The number ofbedrooms can berestricted to anattribute domain of ��0, 1, 2, 3, 4, or 6.

Features can be validated by rules

Objects in the world follow rules when they areplaced or changed. You can use rules to constrainhow the parts of a network are connected or thecardinality of relationships.

A 6-inch pipe can beconnected to a 4-inchpipe only with a properfitting.

Relationships betweenhouses and owners canbe restricted to twoowners per house.

Features can have topology

Many types of features have a precise relationshipthat is characterized as topology.

Parcels of land within a subdivision must adjoineach other exactly, without gaps or overlaps. Thistwo-dimensional graph is called a planar topology.

The lines and devices of a utility network must becontinuously and unambiguously connected. Thisone-dimensional graph is called a geometric network.

Features can have complex behavior

Simple behaviors of features are implemented bychoosing a feature type and topological association,setting up relationships, assigning attribute domains,and specifying validation rules.

More complex behaviors of features can beimplemented by extending a standard feature andwriting software code for a custom feature. Customfeatures permit complex behaviors such as customediting interaction, intrinsic analytical capabilities,and sophisticated cartographic rendering.

Standard featuresand objects

Custom featuresand objects

SMART FEATURES

Features in a geodatabase have a framework ofattributes, geometry, spatial reference, relationships,domains, validation rules, topology, and customobjects. All aspects of this framework, except forcomplex behavior, require no programming.

Features in the geodatabase give you considerablecontrol in modeling the world more naturally.


When you design and build a geodatabase, youemploy a progressive set of steps to add intelligenceto your features. As you have maximized the utilityof each step, you move on to the next to furtherrefine your feature model.

Depending on the requirements and complexity ofyour application, you may find that you may needto employ only a subset of techniques. For example,most applications will not require custom objects.And some applications may not require establishingrelationships among features and objects;topological associations may do the job.

PROGRESSIVELY ADDING INTELLIGENCE

The following are the techniques for tailoring andcustomizing objects. The remainder of this chapterfocuses on these techniques in more detail.

Select feature type and topology

Early in your implementation of a data model, youshould take an inventory of all the types of objectsyou will need to model in your geodatabase. Fromthis inventory, you will establish feature datasets togroup feature classes that are bound by spatialreference, topology, and similar thematic content.

For nonspatial objects, create object classes. Forspatial objects, create simple feature classes withpoint, line, or polygon shapes. For topologicalfeatures, create a graph with topological featureclasses in a common feature dataset.

Set attribution and subtypes

Once you’ve defined the type of object or featureclass, you can add additional fields for theattributes of your object.

Objects and features can have a special attributecalled a subtype. Subtypes are used for majorgroupings of objects and let you express diversityamong similar objects or features without requiringthat you create many object or feature classes.

A subtype for a type of road would let you modeldirt roads, residential roads, and highways, andenforce data integrity specifically for each subtype.Subtypes improve your data integrity throughattribute domains, default values, connectivity rules,and relationship rules.

Define attribute domains and validation rules

An attribute domain is a specified set or range ofvalid attributes. They prevent many simple mistakeswhen you apply a value to an attribute.

A default value applies an expected attribute valuefor a new object. It can streamline data entry byautomatically assigning common attribute values.

Connectivity rules apply to features in a network;they are used to validate whether one type offeature can be correctly connected to another typeof feature.

Establish object relationships

All objects interact with other objects. Importantassociations between objects that cannot becaptured through topological associations can becaptured as relationships. Relationships are stored inrelationship classes and let you control andcustomize how objects and features are created,modified, and removed.

You can define relationship rules on relationshipclasses to further validate exactly how many featuresor objects can be associated with another.

Create custom objects

Object classes, domains, default values, validationrules, and relationships can express the majority ofan object’s desired behavior, but sometimes morecomplex behaviors for drawing, editing, orinspecting objects are needed. The set of ArcInfoobject and feature classes can be extended by aprogrammer to create sophisticated and highlyspecialized objects and features.

SUMMARY

Most customization you will need for objects andfeatures in ArcInfo can be done in the geodatabasedata model without writing any software code.

As a data modeler, one of your main goals is tocapture as much of the natural behavior of yourobjects as possible with this framework. Definingcustom objects and features and writing softwarecode should be necessary only for advancedapplications.

STEPS TO MAKING FEATURES SMART


simple

Set attribution and subtypes

subtype a

subtype b

subtype c

complex

Select featuretype andtopology

Determine the groupings of similar features and objects.

If there are topological associations, select topological feature types.

Otherwise, select simple feature types and specify geometry type.

Create object classes for nonspatial objects.

The spectrum of tailoring features

Define the attributes for object and feature classes.

Decide whether an object class requires subtypes.

Assign names for subtypes.

Assign default values for attributes.

Set up attribute domains for valid values and numeric ranges.

Declare attribute update policies for splitting and merging.

Define relationship rules.

If features are in a network, define connectivity rules.

Declare the types of relationships among object andfeature classes.

Define optional attributes for each relationship type.

Constrain relationship cardinality with relationship rules.

Decide what should happen to a related object when aselected object is changed or removed.

For complex behavior such as custom editing, complex validation, specializeddrawing, or sophisticated analysis, extend the standard object or feature class

types and write software code. Custom objects are necessary only foradvanced data models and applications.

Define attributedomains andvalidation rules

Establish object relationships

Create custom objects


DESIGNING THE GEODATABASE

Because geographic features exist in a rich contextwith topology, spatial reference, and relationships,you have a number of decisions to make whendesigning your geodatabase.

These are the design considerations you should beaware of when you build your geodatabases.

CREATING GEODATABASES

You can work with any number of geodatabases inArcInfo, but in certain situations grouping orsplitting sets of features by geodatabase is better.

These are some reasons to group a set of featuresinto a common geodatabase:

• If a set of objects and features haverelationships, they must be in a commongeodatabase.

• Features that have topological associations mustbe in a common geodatabase.

• If you need to concurrently edit a set of features,they must be in a common geodatabase. Youcan view multiple geodatabases in ArcMap, butyou can edit only one geodatabase at a time.

These are some reasons to separate a set of featuresinto distinct geodatabases:

• If you are working in a large organization,different departments have responsibility forvarious datasets. Geodatabases can be deployedto follow your organizational structure.

• You have the freedom to use any number ofcommercial relationship databases, but eachmust be served through a separate geodatabase.

• If you are working with personal geodatabases,practical size limits may require thematic orspatial partitioning of geodatabases.

ORGANIZING FEATURE DATASETS ANDCLASSES

A geodatabase contains three general types ofclasses: object, feature, and relationship. Theseclasses can either reside in a feature dataset or asstand-alone classes in a geodatabase. These aresome reasons to group classes in a feature dataset:

• If feature classes are topologically related by ageometric network or planar topology, they mustreside within a common feature dataset.

• If you want to enforce a common spatialreference for a set of feature classes, they shouldbe in a common feature dataset.

• You can also arbitrarily group thematicallyrelated classes in a feature dataset.

There is no restriction on the placement ofrelationship classes; they can reside anywherewithin a geodatabase and represent origin anddestination classes throughout the geodatabase. Ifthe origin and destination classes of a relationshipclass are in a common feature dataset, that is a goodlocation for the relationship class, but it is notrequired.

APPLYING SUBTYPES

One of the most important design decisions youwill make is whether a group of related featuresshould constitute a set of feature classes or a singlefeature class with features segregated by subtype.

A subtype is a lightweight classification of features(or objects) within a feature (or object) class. Thekey motivation for using subtypes is performance. Ageodatabase with one or two dozen feature classeswill perform better than a geodatabase with manydozens of feature classes.

Subtypes let you control specific behavior for a setof features in a feature class through attribute rules,default values, connectivity rules, and relationshiprules. Whenever possible, your first preferenceshould be to use subtypes to differentiate groups ofrelated features.

These are some reasons why it is sometimesnecessary to split groups of related features intodistinct feature classes:

• When each group of related features requiresdistinct custom behavior.

• When the set of feature attributes is different. (Allfeatures in a feature class have the same set ofattributes.)

• When you require different access privileges foreach group of features.

• When some features are to be accessed throughversions and some are not.


A geodatabase contains feature datasets,stand-alone object classes, feature classes,relationship classes, and attribute domains.

A feature dataset contains object, feature, andrelationship classes, a spatial reference, and

geometric networks.

Object classes store nonspatial entities and havesubtypes and attribute rules. (In the geodatabase,

“object class” and “table” are synonyms.)

Feature classes store simple or topological featuresand have subtypes and attribute rules. A stand-

alone feature class can have only simple features.

Relationship classes store relationships, may haveattributes, and have associated relationship rules.

All feature classes in a feature dataset share thesame spatial reference. Stand-alone feature classes

also have a spatial reference.

All feature classes that participate in a geometricnetwork must be in a common feature dataset. Each

geometric network has a set of connectivity rules.

Any feature classes that you want to perform two-dimensional topological editing on must be in the

same feature dataset. This ensemble of featureclasses is called a planar topology.

Note: At the initial release of ArcInfo 8, planartopologies are not yet persistently stored in the

geodatabase, but are dynamically defined whenperforming topological editing in ArcMap.

Attribute domains are stored directly in thegeodatabase, ready to be applied to any attribute of

any object or feature class in the geodatabase.When it is applied, it becomes an attribute rule for

that object or feature class.

Structure of features and objects

Geodatabase

Feature datasets

Spatial reference

Geometric networks,connectivity rules

Planar topologies

Attribute domains

Feature classes,� subtypes, attribute rules

Object classes, subtypes, attribute rules

Relationship classes, relationship rules

This conceptual illustration shows how features and objects are structured in a geodatabase. At theend of this chapter is a UML diagram of the geodatabase structure from a programmer’s perspective.

can be

inside or

outside of

feature

datasets


Tables are the repository of objects and theirattributes. A table stores attributes for objects thatare reasonably similar to one another and have thesame set of attributes. For example, a table couldstore records for persons, buildings, and roads.

TABLES AND ROWS

A table is organized into rows and columns.

A row is the fundamental unit of information in atable and comprises a set of properties for an object.All the rows in a table must have the same set ofproperty definitions.

A column represents all the attributes of one type.The value of a column for a given row is called anattribute. The definition of a column—its name andwhether the column is formatted to store an objectidentifier, geometry, real numeric value, integernumeric value, or character string—is called a field.

Types of tables

In a geodatabase, tables can store nonspatialobjects, spatial objects, and relationships.

A table that stores nonspatial objects is called anobject class. It has a special field for subtypes.

A table that stores spatial objects is called a featureclass. Simple feature classes have two predefinedfields: a feature ID and a geometry field. Annotationfeature classes and network feature classes haveadditional predefined fields.

A table that stores relationships is called arelationship class. It can have any number of customfields to represent the attributes of the relationship.Not all relationship classes are implemented astables. If a relationship class is not attributed anddoes not have a cardinality of many-to-many, it isstored as a set of foreign keys on the feature orobject classes. Attributed relationships or many-to-many relationships are stored in tables.

Fields in a geodatabase

Attributes can express several qualities of an object.These are some common types of attributes:

• An attribute can designate a coded value for aclassification.

• An attribute can be descriptive text thatcharacterizes a feature or gives its name.

• An attribute can characterize a real numericvalue that is measured or calculated, such asdistance or flow.

• An attribute can represent a counted value, suchas the number of associated parts.

• An attribute can specify a unique identifier thatreferences a row in another table.

A table in a geodatabase can support these andother types of attributes with these field types: float,double, short integer, long integer, text, date, object ID,and binary large object (BLOB).

Predefined and custom fields

There are two sets of fields in a table: predefinedfields for uniquely identifying objects and storingfeature shapes, and custom fields for definingadditional attributes of features. Predefined fieldsand custom fields coexist in the same feature classtable.

Predefined fields are managed by ArcInfo andshould never be modified through another databaseapplication.

Custom fields implement the various types ofattributes needed to realize the properties of yourfeatures. In the diagram, custom fields describe roadtype, surface, width, lanes, and name. You can addany number of custom fields.

Attribute and spatial indexes

You can create attribute indexes on fields to makequery performance quicker. In ArcCatalog, you cancreate indexes on one or several attributes in a tableand you can add and remove indexes at any time.Be careful with indexes—performance diminisheswhen you have defined an excessive number ofindexes.

ArcInfo automatically creates spatial indexes onfeature classes. It determines and applies anoptimum grid size for you. To optimize certainfeature classes, particularly when feature size variesconsiderably, you can define up to three grid sizesfor best retrieval of spatial data.

STORING DATA IN TABLES


A feature class with line geometryhas several predefined fields: aunique feature ID, a geometry-

tracking field to record the featurelength, and a geometry, which is the

shape of the feature.

The other fields shown are customfields. Some examples of custom

attributes are coded values for roadtype, a descriptive string for surface

type, a continuous value for roadwidth, a discrete numeric value for

the number of lanes, and text fornames.

All attributes are based on one ofthese types: float, double, short

integer, long integer, text, and date.

A table is a set of rows.

A row is a set of attributes.

Columns represent all attributes of the same type in a table.

A field is a description of a column.

All rows in a table have the same set of fields.

You can also use anattribute as an identifierto rows in another table.

This related table iscalled a lookup table.

2 arterial or collector roads

1 divided highway

3 major roads

4 residential streets

code description

5 unpaved roads

predefinedfields

customfields

fid geom shp_len type surface width lanes name

103 2321.8 3 asphalt 75.9 4 Caitlin County Road

101 4507.2 2 asphalt 85.3 4 Old Taos Highway

102 3401.1 1 concrete 45.1 2 Calle Mejia

104 689.2 5 gravel 35.2 2 Max Daniel Road

object identifiergeometrygeometry-tracking fieldcoded valuedescriptive stringcontinuous numeric valuediscrete numeric valuename

Tables, objects, and attributes

row

column

attribute

An object class is a database table in a geodatabase. In addition to the basic functions on tablesand their rows and columns, you can apply some of the functions of feature classes such assubtypes, attribute domains, default values, and relationships.

A feature class is an object class with a geometry field to keep the shape of features.


You can think of a GIS as an extension of databasetechnology that stores, manages, and updatesspatial information. Features are spatial objects andmuch of the functionality inside ArcInfo involvesthe display, query, and editing of features.

A feature class has a special field that represents theshape and location of features. This field is calledshape and is of the field type geometry. All featuresin a feature class have the same type of geometry.

FEATURES AND GEOMETRY

A shape field of a feature class can be one of thefollowing types of geometry: point, multipoint,polyline, or polygon.

A feature with point shape has a single x,y or x,y,zcoordinate value. A feature with multipoint shapehas a number of x,y or x,y,z coordinate values.There is no order implied in the set of coordinatesin a multipoint shape.

A feature with polyline shape has one or morepaths. A path is a connected collection of segments,each of which can be one of these types ofparametric curves: line, circular arc, elliptical arc,or Bézier curve. An optional z value (commonly anelevation) or m value (measurement distance) canbe associated with a feature with polyline geometry.

A feature with a polygon shape has one or morerings. A ring is a connected, closed, andnonintersecting set of segments. Each segment canbe of type line, circular arc, elliptical arc, or Béziercurve. A ring cannot intersect itself, but canintersect other rings in a polygon. Rings in apolygon can touch at any number of points. Anoptional z value (commonly an elevation) can beassociated with a feature with polygon geometry.

An important change in ArcInfo 8 from previousreleases is that single- and multiple-part geometriesare now integrated in the same feature class. In thecoverage data model, single-part lines were kept inarc feature classes, multipart lines in route featureclasses, single-part polygons in polygon featureclasses, and multipart polygons in region featureclasses. Single-part and multipart geometries are nolonger stored separately in the geodatabase.

THE SHAPE AND EXTENT OF FEATURES

Another important change is support for parametricsegments: circular arcs, elliptical arcs, and Béziercurves. These types let you more accuratelyrepresent the feature shapes and are especiallyimportant for civil engineering applications.

A specific feature in a feature class can contain anull geometry. The data modeler may use nullgeometry to represent objects that are sometimesrepresented as explicit features and sometimes asimplicit features within composite objects.

Chapter 6, “The shape of features,” contains moreinformation about the geometry of feature shapes.

FEATURES AND SPATIAL REFERENCE

The geometry of features is stored as a structuredset of x and y coordinates, with optional z and mvalues. These coordinates are related to the shape ofthe earth through a spatial reference.

One part of the spatial reference is the coordinatesystem, which defines the mathematical projectionof a planar area to the roughly spherical shape ofthe earth, called a geoid.

The other part of a spatial reference defines how aset of coordinates is related to its storage in ageodatabase as integer values. The geodatabaseuses integers internally to prevent ambiguities whencomparing locations and applying spatial operators.When you use ArcInfo, these integer values areconverted to map units and you are not aware ofthem except when you need to define the spatialdomain and scale of a spatial reference.

The spatial domain consists of minimum andmaximum values for x and y, and optionally z andm. The scale defines how many integers correspondto a map unit. If the scale is 1,000, then themaximum precision is 1/1,000 of a map unit.

There is an inverse relationship between scale andspatial domain. If you select a very high value forthe scale, the allowable spatial domain is restricted.As a rule of thumb, the product of the scale andthe greatest range of the spatial domain cannotexceed two billion (or 2 to the 31st power).

For more information about spatial references, readthe ESRI Press book Understanding Map Projections.


Geodatabase

Feature dataset

Feature class

Feature class

Feature class

Spatial reference

Spatial reference

Spatial reference

Feature geometry

Points Lines Polygons

Path Ring

Line Circular arc Elliptical arc Bézier curve

Segments

Single-part polylinePoint

Multipart polylineMultipoint

Single-part polygon

Multipart polygon

Features in a geodatabase haveone of four types of geometry: point,

multipoint, polyline, and polygon.

Polylines are comprised of one ormany paths. Polygons are

comprised of one or many rings.

A path is a simple, connected series of any ofthe four types of segments. A path cannot

intersect itself.

A ring is a path that is closed.

Four types of segments are presentlysupported in a geodatabase. Advanceddevelopers can introduce new types of

segments such as spirals.

A spatial reference has three main components:

A coordinate system defines a map projection and itsparameters.

A spatial domain constrains the range of x and y values,and optionally z and m values.

A scale defines how many integer units correspond to amapping unit, and defines the coordinate precision.

A spatial reference is associated with a feature class. Iffeature classes are organized within a feature dataset, all ofthose feature classes share the same spatial reference.

A geodatabase can have many spatial references, one foreach feature dataset and each stand-alone feature class.

Once a spatial reference is assigned to a feature class orfeature dataset, you can update the coordinate system, butnot the spatial domain or scale.

Spatial reference

scalenumber of internal integer units to onemap unit

range of x coordinates

spatial domainminimum and maximum x, y, z, andm values

coordinate systemdefined by projection (cylindrical,conic, planar, other) with parameters

rang

e of

y c

oord

inat

es


An attribute is a quality of an object. An attributefor a feature could be its size, density, name, flow,date of installation, or population.

Each object or feature in a GIS dataset has anumber of attribute values, which are kept as rowsin a database table. The attributes collectivelyrepresent the important qualities of that feature typefor your application. Chapter 2, “How mapsinform,” discussed how attributes can be shown ona map. ArcMap supports a rich set of drawingmethods to present attributes.

Egyptnumericattributes

classifiedattributes

typeattributes

descriptiveattributes

TYPES OF ATTRIBUTES

A number of types of attributes can be associatedwith a feature in a geodatabase. The following aresome examples drawn from roads.

Continuous numeric values with floats anddoubles

An attribute can contain a real numeric value, suchas 2.7. These attributes are continuous data that ismeasured or calculated, such as distance or flow.A road has numeric attributes for its length, routemeasure, or width.

Discrete numeric values with shorts and integers

Some attributes represent numeric values that iscounted, not continuous. They are usually positiveinteger values, but can be negative in some cases. Aroad has integer attributes for the number of lanes.

Coded values with shorts, integers, and text

An attribute can be a coded value. The value itselfis not meaningful, but it references other attributesthat are uniquely tied to the coded value. A road

feature may have an attribute designating road typesin this fashion:

1—divided highway

2—arterial or collector roads

3—major roads

The advantage of coded values is that they take uplittle space in a table; their disadvantage is that youare one step removed from a meaningful attribute.

Descriptions with text

An attribute can be a descriptive string thatcharacterizes a feature or gives its name. A roadfeature might have these allowable values for itssurface type: “asphalt,” “concrete,” and “gravel.”

A road feature would likely have an attribute for itsname, such as “West Manhattan Avenue.” This namemight be kept in its entirety, or might be brokendown into a road prefix (“West”), road name(“Manhattan”), and road suffix (“Avenue”). Exactlyhow a road name is represented is important if youare performing address matching in your GIS.

Time values with dates

Important events on objects can have a date valueassigned that records a time. This is the temporaldimension of your GIS. When a road’s pavement isreplaced or maintained, a road inventory databasecan mark the time of that event with a date value.

Object identifiers

The power of a relational database is realized whenrelationships are made between rows or featuresfrom different tables. A feature attribute can be anidentifier that references a feature or row in anothertable. For a road feature, an object identifier mightreference an annotation with the road’s name or amaintenance record of that road. When you createnonattributed relationships or annotation features,object identifiers are inserted for you.

Multimedia with BLOBs

Tables in a GIS can contain a BLOB (binary largeobject) column. This enables you to integrate othermedia such as video, images, or sound. A section ofroad can be associated with roadside photographsstored in a BLOB column in a table.

ATTRIBUTES: QUALITIES OF AN OBJECT


Egypt

Text showsnames and

other qualitiesof features.

Description

-14

5664

143

short integer

long integer

2.310.0

-4.78.63

float

double

Arkansastext

SeaBlvd

Red45th

A short integer value containsone sign bit and 15 binary bitswith a range of approximately–32 thousand to 32 thousand.

A long integer value containsone sign bit and 31 binary bitswith a range of approximately–2 billion to 2 billion.

A float value contains one signbit, seven exponent bits, and24 mantissa bits.

A double value contains onesign bit, seven exponent bits,and 56 mantissa bits.

Text values contain any numberof characters. Each character isstored in a byte (8 bits). All textvalues in a field have the samenumber of characters withtrailing blanks.

A B C D E F G H I J

Graduatedsymbols

Any type ofnumeric valuecan be drawn

with graduatedsymbols,

which vary inproportion to a

value.

Classifiedvalues

A classificationis a statisticalsubdividing ofthe numeric

values of a setof objects.Classifiedvalues aredrawn with

color ramps.

Unique values

Someattributesrepresent

categories ortypes. A

random set ofsymbols isapplied to

each uniquevalue.

Attribute types Applications

12/1/611/30/947/16/97

dateThe date value is translatedinto the current day and timein the local time zone.Date values are based on a

standard time format.

635432689764

object ID An object ID value is a longunique identifier generated ingeodatabases.

Object IDs are used fordatabase joins andestablishing relationshipsbetween objects.

BLOB BLOB values contain complexobjects like images and video.

BLOB values let you add anykind of multimedia content toyour geodatabase tables.


ADDING SIMPLE BEHAVIOR WITH SUBTYPES

Every user’s goal when adding or editing objectsand features in a GIS database is to eliminate orminimize data entry error. For many modelers, thisis the most important aspect of designing ageodatabase.

The easiest way to add intelligent behavior thatverifies the integrity of features and objects is to:

• Apply constraints for updating attributes

• Define validation rules for how features arereferenced or located with respect to each other

Further, you would like a fine level of control sothat you can define behavior that discriminatesbetween subgroups of feature classes, or subtypes.

Subtypes

Objects in an object class and features in a featureclass may be further subdivided into subtypes.

A subtype is a special attribute that lets you assigndistinct simple behavior for different classificationsof your objects or features. All subtypes of a classshare the same set of attributes.

The motivation for defining subtypes of an objectclass is to introduce a lightweight subdivision of anobject class that adds these capabilities:

• You can name subtypes to describe each memberof a classification of your objects.

• You can define distinct attribute domains for eachfield in a subtype.

• You can define distinct default values for eachfield in a subtype.

• You can prescribe the types of relationships thatare possible between the objects in a subtypeand objects in another subtype in the same ordifferent object class.

• If you write some software code, you can alsoadd custom rules for subtypes of object andfeature classes.

An object class does not have to contain subtypes. Ifnone are defined, you can still set attribute domains,default values, and rules—but on the object orfeature class as a whole instead of on a subtype.

Attribute domains

Constraints on attributes are called attributedomains. For numeric attributes, you can set arange domain that constrains the value to betweenprescribed minimum and maximum values. Anexample is to constrain the price of one hectare ofland to between 10,000 and 1,000,000 euros.

For all attribute types, except object IDs and BLOBs,you can set a coded value domain, which is adefined set of valid values. An example of codedvalue would be a list of geologic strata types suchas “Pre-Cambrian,” “Jurassic,” and “Cretaceous.”With a coded value domain, you can ensure thatthe attribute definitely has one of the expectedvalues.

When editing features and objects in ArcMap, youcan enter features with invalid values, but you canvalidate your work at any time. Invalid attributevalues are highlighted for editing.

Validation rules

Validation rules control feature and attributeintegrity. The types of validation rules are attributerules, connectivity rules, and relationship rules.

An attribute rule is an attribute domain applied to asuptype of a class. An example of an attribute ruleis that the field named DIAMETER can representonly pipes that are 10, 15, 25, or 50 centimeters indiameter.

A connectivity rule specifies the valid pairs ofattribute values for subtypes for connected networkfeatures. For example, an electric line with phaseABC may be connected to a downstream line withphase AC. The types of connectivity rules are edge–junction rule, edge–edge rule, default junction type,and edge–junction cardinality.

A relationship rule constrains the cardinality of arelationship between an origin class and destinationclass. The four basic cardinalities are one-to-one,one-to-many, many-to-one, and many-to-many. Witha relationship rule, you can create specializedcardinalities, such as a state has exactly twosenators; a parcel of land can have no, one, or twoowners; or a pole can have no, one, two, or threetransformers mounted on it.


For each subtype in your class, you can define simple object behavior withdefault values, attribute domains, connectivity rules, and relationship rules.

A subtype is stored as an integer value with a descriptive name such as“asphalt.”�

Each subtype in an object class or feature class has a default value for newattributes, domains of valid attributes, rules to validate how features connector relate, and the type of relationship possible for a new object.

fid geom subtype width lanes name

103 asphalt 75.9 4 Calle Petra

101 asphalt 85.3 4 Chimayo Highway

102 concrete 45.1 2 Acequia de Isabel

104 gravel 35.2 2 Maximilian Road

Simple behavior with subtypesAn object class or feature class can have a special type offield called a subtype. A subtype is used for what youregard as the most significant classification of the objectsin your object class.

Subtypes help you preserve the integrity of your data.

Features sorted by subtype Realizing simple behaviors

Validation rulesA concrete roadcan connect toan asphalt road

but not to agravel road.

A two-laneasphalt road

can onlyconnect to

another two-lane road.

A gravel roadcannot directlyconnect to a

freeway.

Valid widths are55, 65, and 75.Valid suffixesare “Highway”�

and “Interstate.”�

Valid widths are30, 35, 40, and45. Valid lane

counts are 1, 2,and 4.

Valid widths are15, 20, and 25.Valid suffixes

are “Road” and“Lane.”�

A new concreteroad is given adefault value of

four lanes.

A new asphaltroad is given adefault width of

35 feet.

A new gravelroad is given adefault width of

15 feet.

Two concreteroadways can beassociated with a

highway route.

An asphalt roadcan be relatedwith bridges ortunnel crosses.

A gravel roadcannot have more

than four roadsegments at an

intersection.

103 concrete 75 4 NM Highway 14

102 concrete 65 4 US Highway 285

104 concrete 75 4 US Interstate 25

103 asphalt 40 2 Acequia Wier

101 asphalt 45 2 Grant Paige Ave

102 asphalt 35 2 Shakedown Street

104 asphalt 45 2 Hart Alley

103 gravel 20 1 McKernan Lane

102 gravel 15 1 Lesh Ranch Road

104 gravel 15 1 Kreutzman Road

101 gravel 25 2 Garcia Road

fid geom subtype width ln name

Connectivityrules

Attributedomains

Defaultvalues

Relationshiprules

A split highwayretains allhighway

designations.

A mergedasphalt road

takes a defaultvalue for lanes.

A split gravelroad retains its

width.

Split/Mergepolicy

Roads with subtypes


VALIDATING ATTRIBUTES

An attribute domain is a constraint on attributevalues in feature and object classes. This constraintcan be a range of numeric values or a list of validvalues.

Attribute domains are organized in a geodatabaseand are ready for any object or feature class to use.When an attribute domain is applied to a particularattribute in a subtype, it becomes an attribute rule.

The following sections outline the components ofattribute domains.

Range domains

To prevent data entry error, a range domain canconstrain the value of any numeric attribute in anyobject or feature class to minimum and maximumvalues. An example of a range domain is thepressure in a pipe is expected to be between 2,000and 14,000.

A range domain can be applied to short-integer,long-integer, float, double, and date attribute types.

Coded value domains

Many attributes are classifications of features. Forexample, a land-use type can be constrained to alist of values, such as “residential,” “commercial,”and “park.” You can update the list of valid valuesin a coded value domain at any time.

A coded value domain can be applied to text,short-integer, long-integer, float, double, and dateattribute types.

Default values

During data entry, it is frequently the case that for acertain attribute, one value is most commonlyexpected. Default values apply the expected valuefor a subtype in an object class when the feature iscreated, split, or merged. An example of a defaultvalue is applying “residential” as the default land-use classification of a new or split land parcel.

A default value can be applied to text, short-integer,long-integer, float, double, and date attribute types.

Splitting features

Once you have set a range or coded value domain,you can refine that domain by declaring whathappens when features are split.

A land parcel split is a common scenario. Whenone piece of land is split into two, you might valuethe new parcels based on the proportion of theirsizes. Or, you may want to apply an attribute valueto both split parcels. You might also apply a defaultvalue to a new attribute.

The split policies are:

• Default value—A default value is applied toattributes of both split features.

• Duplicate—The attributes of both split featuresare identical to the value of the original featureattributes.

• Geometry ratio—You can define attributes ofsplit features to be the proportional value of thesplit areas or lengths.

A split policy can be applied to text, short-integer,long-integer, float, double, and date attribute types.

Merging features

You can further refine an attribute domain by settingwhat happens to attributes when two objects aremerged into one.

The merge policies are:

• Default value—A default value is applied to themerged features.

• Sum values—Two numeric attribute values aresummed for the attribute of the merged feature.

• Weighted average—The attribute of the mergedfeature is the weighted average of the values ofthe attribute from the original features.

A merge policy can be applied to text, short-integer,long-integer, float, double, and date attribute types.An example of a merge policy is combining twocrop yield values into one for a merged parcel offarmland.


Controlling attributes with domains

Steps to setting attribute domains

short integerlong integer

floatdoubledate

text

1select an attribute

range minimum value maximum value

coded value

domain type

2set a domain type

Text can only have a coded valuedomain. All other attribute types can have

a coded value or range domain.

duplicatedefault valuegeometric ratio

duplicatedefault value

default valuesum valuesweighted average

default value

split policy merge policy3set split and merge policies

Based on the domain type, these are thevalid split and merge policies available.

Attribute domains, default values, and split and merge policies are techniques thatmake it easy to validate attributes of features and objects.

merge policy sum values weighted average

default value forattribute is applied tothe merged feature

numericattribute is

summed

numeric value is weightedaverage of attribute from

the two features

Riverside

district

Lakeview 24000 35

45000 47

yield % harvested

district

Montane 69000 43

yield % harvested

default value

Merging features

split policy duplicate geometry ratiodefault value

attribute of originalfeature is duplicated

in split features

numeric attribute issubdivided by ratio of

split area or length

default value isapplied to split

features

G Gould

owner

G Gould 14000R-1

11000R-1

valueland zoning

owner

G Gould 25000R-4

valueland zoning

Splitting features

2,000 to 14,000 residential, commercial, park

pressure land use

A range domain specifies that an attributevalue must be between a specified minimum

and maximum value.

A range domain can be assigned to any typeof attribute except text, object ID, or BLOB.

Attribute domainsRange domain Coded value domain Default value

R-1

land zoning

A coded value domain is a set of valid valuesfor an attribute.

A coded value domain can be assigned to anytype of attribute except object ID or BLOB.

A default value can be applied to an attributewhenever an object or feature is created.

A default value can also be applied when afeature is split or merged.


Objects in the world have relationships with otherobjects. Some objects have a spatial extent, such asroads. Other objects do not have a fixed spatialextent, such as people.

Some examples of relationships among objects arethat a parcel of land can be associated with anowner, a land-use zone, an annotation with a lotnumber, or a building.

It is desirable to keep track of these relationships sothat when one object is modified, the relatedobjects can react appropriately. For example, whena utility pole is removed, the attached transformersand other equipment are removed as well.

The geodatabase provides a framework to explicitlydefine relationships among features and objects.ArcInfo includes the functionality to manage theserelationships and ensure feature integrity.

WHEN TO USE RELATIONSHIPS

ArcInfo has three ways to define relationshipsamong features: topological, spatial, and general.

What is connected or adjacent

Topological relationships are built into the datawhen you create a geometric network or planartopology. These relationships can quickly findneighboring polygons and traversed lines. They aremanaged for you through the topologicalenvironment of the ArcMap Editor.

What is spatially related

ArcInfo contains a rich set of spatial operations thatcan determine whether one feature touches,coincides with, overlaps, is inside of, or is outside

of another feature. For example, you might want todetermine which building footprints are inside of aland parcel.

General relationships

transformermeter

parcel

owner

General relationships are explicitly definedrelationships that form a persistent tie between afeature or object from an origin class to a feature orobject in a destination class.

The data modeler can explicitly model generalrelationships between objects that cannot necessarilybe inferred from their geometry or topology.

These are some uses of general relationships, againillustrated through examples of a road feature:

• A one-to-one relationship might be between aroad feature and an associated row in anexternal table, such as road maintenance data.

• A one-to-many relationship might be between aroad feature and a set of events, such as trafficaccidents.

• A many-to-many relationship might be betweenmany road features and many highwayconstruction work orders.

Any of these relationships can be establishedbetween the features and objects in a geodatabase.

RELATIONSHIPS AND RELATIONSHIP CLASSES

A relationship is an association between two objects.These objects can be nonspatial (objects) or spatial(features). Besides identifying the associated objects,relationships can have additional properties.

Relationships are organized into relationship classes.Each relationship in a relationship class has thesame origin class and destination class. Any objectclass may participate in many relationship classes.With relationships, the geodatabase ensuresreferential integrity between objects as they arecreated, modified, or deleted.

RELATIONSHIPS AMONG OBJECTS


RelationshipsThe geodatabase uses relationships between objects to maintain the referential integrity of

objects when they are deleted or moved. Related objects can issue notifications to each otherwhen there is a change.

Cardinality of relationships

relationshipclass

62

65

69

21

23

29

11

13

17

One-to-one relationships

62

65

69

21, 23

23

27, 29

11

13

17

Many-to-one relationships

62, 64

65

64, 68, 69

21

23

29

11

13

17

One-to-many relationships

62

64, 65

64, 69

21, 23

23

27, 29

11

13

17

Many-to-many relationships

originclass

21

23

27

29

21

23

27

29

62

64

65

68

69

21

23

27

29

21

23

27

29

62

64

65

68

69

destinationclass

62

64

65

68

69

62

64

65

68

69

relationshipclass

originclass

destinationclass

relationshipclass

origin class

destinationclass

relationshipclass

originclass

destinationclass

There are four basic cardinalities of relationships. The cardinality affects whether a relationshipcan be simple or composite.

Pole–transformer relationship examplehas attachments

is attached to

103 wood 35

101 concrete 25

102 wood 35

104 steel 45

Pole feature class

fid geom subtype height fid geom poleID kVA

308 102 25

301 101 50

305 102 25

311 102 25

Transformer feature class

origin class destination classpath labels

relationships

The pole feature class is considered the originclass because transformers are mounted onpoles.

A pole–transformer relationship is considered acomposite relationship because the placementand lifespan of a transformer is affected by thelifetime of the pole.

One pole can be mounted with zero, one, two, orthree transformers, so a relationship rule wouldbe defined to enforce this cardinality constraint.

A composite relationship ensures that when an object from theorigin class is moved or deleted, the related object from thedestination class is also moved or deleted.

select a pole and... ...transformer moves with pole

poletransformermeterprimary linesecondary line

The associationbetween utilitypoles and electrictransformers is anexample of a one-to-manyrelationship.

pole

transformers


Path labels

A relationship class has a forward path label and abackward path label. These labels describe therelationship and are used when displaying objectrelationships. Examples of path labels are“manages” and “managed by.”

Cardinality

A relationship class has cardinality, which restrictsthe number of relationships that can be formedbetween an origin object class and destinationobject class. Examples of cardinalities are one-to-one, one-to-many, many-to-one, and many-to-many.

A relationship class can be simple or composite.

A simple relationship is a peer-to-peer relationshipwhere related objects can exist independently.

A composite relationship is a one-to-manyrelationship between a composite object from anorigin class and part objects from a destination class.A composite relationship must have one-to-manycardinality, and you can use relationship rules toenforce this.

When an object from an origin class is deleted, therelated objects in the destination class must also bedeleted.

Part objects can be created independently of thecomposite object, but they must be deleted whenthe composite object is deleted. Part objects of acomposite object can be deleted and replaced bynew composite objects.

Notifications

A notification is a message passed in ArcInfo whena significant event occurs, such as an edit ordeletion. Notifications are the mechanism thatmanages the lifespans of part objects based on thewhole object in a composite relationship.

A relationship class may be used to propagatestandard notifications between related objects. Thenotification direction property specifies these fournotification options:

• No notifications are propagated.

• A notification is issued to the destination objectonly when the origin object is changed.

• A notification is issued to the origin object onlywhen the destination object is changed.

• A notification is issued when either the origin ordestination object is changed.

Attributed relationship classes

When ArcInfo creates a one-to-one or one-to-manyrelationship class, it implements it as insertedforeign keys on the origin and destination class. Arelationship table is not built with these cardinalitiesunless you specify that you want to add attributes torelationships.

When ArcInfo creates a many-to-many relationshipclass, a relationship table must be built because theforeign keys cannot unambiguously record all of therelationships.

A relationship table has a row for each relationship.You can optionally add attributes to relationships.These attributes can be any qualities that describean event that binds the related objects, such as awork order or accident report.

An example of an attributed relationship is betweenan owner and a piece of land. This relationship canrepresent a transaction in land ownership and theattributes could recite the facts recorded in the titledocument.

simple

simple orcomposite

simple orcomposite

managed with foreign keys orrelationship table if

relationships are attributed

one-to-many

many-to-many

one-to-one

relationship table is necessary

Cardinality Control Implementation

managed with foreign keys orrelationship table if

relationships are attributed


ANNOTATION AND ANNOTATION CLASSES

An annotation is a type of feature that provides atextual description of a place or feature.

An annotation feature class is a feature class thatcontains annotation. All annotation in an annotationfeature class has the same set of attributes.

Feature-linked annotation

Most features on a map have annotation.Annotation is usually a place name, but can also beany attribute of the feature.

Annotation can be closely bound to features bydefining a composite relationship between a featureclass and annotation class. This is called feature-linked annotation.

The annotation feature class includes propertiessuch as the field from which text labels are derived,the type of symbol, and other attributes.

When a feature in the composite feature class iscreated, a notification is sent to the annotationfeature class, enabling automated placement ofannotation. When a composite feature is deleted,the associated annotation is deleted. When changesare made to the composite feature, notifications areforwarded to the annotation class using thestandard complex relationship notifications.

Simple annotation

Maps also have annotation that is not linked to afeature. Simple annotation can be used for:

• Map graticule information such as coordinate orlatitude values

• Large or indeterminate geographic entities thatare not represented by a single feature

• Any free-form text labeling on the map

Annotation

featureclass

annotationfeature class

compositerelationship

class

Topanga Canyon

Sacramento

Placerville

Feature-linked annotation Simple annotation

annotationfeature class

411 Sierra Nevada Range

413 Galisteo Basin

417 Death Valley

94 Sacramento

95 Topanga Canyon

92 Placerville41

43

4749

Feature-linked annotation is implementedas a composite relationship between afeature class and an annotation featureclass. The feature controls the lifespan ofthe annotation.

Simple annotation iskept in an annotationfeature class that isnot bound in arelationship withfeatures.

When a feature with feature-linked annotation is removed, theannotation is removed as well. When an attribute is changed,the text displayed by the annotation feature class is updated.

Simple annotation has no relationship toattributes of any other feature.

Sierra Nevada Range

Galisteo Basin

Death Valley

There are two types of annotation. Feature-linked annotation provides dynamic labeling offeatures. Simple annotation allows free-form placement of text on the map.

9294

95

41

43

47

21

23

27


The highest form of customization in ArcInfo iscreating custom features. Certain complex behaviorscannot be expressed with rules. These includecustom editing, complex validation, specializeddrawing, and sophisticated analysis.

If you are not a programmer, you can skip theremainder of the chapter. These pages discuss theconcepts of programming complex behavior intocustom features and the programmer’s view of thegeodatabase data access objects.

Object fundamentals

An object represents an entity such as a house, lake,or customer. An object is stored as a row and hasbehavior expressed as one or more sets of methods.

A method has a name, a set of input parametersand output parameters, and a return type. A set ofmethods is called a software interface (or simply,interface).

The developer of an object writes software codethat provides implementation for methods. TheArcInfo system invokes these methods to expressthe complex behaviors of objects.

An object class represents a set of objects that sharethe same type, such as house, lake, or customer. Thebehavior of an object class is implemented with abehavior class that is stored in a dynamic-linkedlibrary (DLL) in conformance with the MicrosoftComponent Object Model (COM) architecture.

The ArcInfo object classes

ArcInfo provides a hierarchy of object classes readyfor use. These are object, feature, simple junctionfeature, complex junction feature, simple edge feature,and complex edge feature.

The ArcInfo object classes include a number ofpredefined fields such as object identifiers andgeometries. These fields define the requiredproperties of the objects in these classes. The datamodeler can add additional custom fields.

Each of these object classes implements a set ofinterfaces. Each interface contains a set of relatedmethods for actions such as storing, editing,drawing, querying, and validating objects.

An ArcInfo object class provides defaultimplementation for each of its interfaces. Thedeveloper can selectively override the defaultimplementation for an interface.

Custom features

The developer can customize object classes byextending the ArcInfo object classes. ArcInfoincludes a CASE (computer-aided softwareengineering) tool framework to graphically extendthe standard ArcInfo object classes. This frameworkautomates the schema generation for the newobject classes and generates source code forbehavior class templates.

The developer implements new interfaces or usesexisting interfaces to model specialized behavior.The developer can also override existing softwareinterfaces inherited from standard ArcInfo classes.

Type inheritance

Features can be specialized. You can create a newtype of custom feature that has all the attributes andbehavior of another, but adds new attributes andbehaviors. For example, a state highway is a type ofroad. This is called type inheritance.

Features are COM objects. COM is an infrastructureto build software components on an inheritancemodel based on interfaces. An interface is acontract between server and client to providespecified services, such as drawing or selection.

In the illustration to the right, the Pole customfeature implements four interfaces: IPole,IFeatureDraw, IFeature, and IRow.

Internals of a custom object

A custom object is a combination of a database tableand code compiled to a DLL.

A custom feature is implemented internally inArcInfo as a feature class table, as a behavior classstored in a DLL, and as a globally unique identifierin the Windows Registry that binds the feature classand its behavior class.

EXTENDING OBJECT CLASSES


Extending the object modelYou can apply specialized behaviors of objects and features by writing software code thatconforms to the interfaces and conventions of ArcObjects, the components inside ArcInfo.

CLSID {AB9-F00} "C:\...\Electric Utils.dll"

Registry key

Feature classes tableElectricUtils.dll

Name ClassID Type

Pole {AB9-F00} SimplePole

Internal implementationThe behavior for custom objects is implemented with COM interfaces in

DLLs. The class table and DLL are linked through the Registry.

Row

Feature

Pole

IRow

IFeature

IFeatureDraw

IPole

ArcInfo

Custom

A pole agrees to act exactlylike a feature. It adds itscustom behavior throughthe IPole interface.

A feature is a type of rowthat stores geometry in atable column. It addsbehavior to interact withand draw features.

A row is a record in adatabase table. You cancreate custom objectsbased on rows.

Type inheritanceA custom object is created by applying type

inheritance using COM interfaces.

Network-Feature

Feature

Circuit Pole

Diagramkey

Standardfeature class

Custom class

Type inheritance

Customabstract class

Standardabstract class

Type inheritance denotes an “is a type of”�relationship. In this key, NetworkFeature, Circuit,and Pole are types of a Feature.

Row

Junction-Feature EdgeFeature

Network-Feature

Feature

Complex-Edge-

Feature

Simple-Edge-

Feature



Standard ArcInfo objects Custom objects

Pole GuyWire

Customer

Segment

Primary-Segment

Secondary-Segment

TransformerMeter Switchgear

Poles and Guywires are structural features that do notparticipate in the network, so they are derived fromthe ArcInfo Feature object.

A Customer has relationshipswith other objects, but is not a feature on the map, so it is derived from the ArcInfo Row �object.�

PrimarySegment and SecondarySegmentare types of Segment, which in turn is atype of SimpleEdgeFeature.

Meters and Transformers aresimple features at network

junctions, so they are derived fromthe SimpleJunctionFeature class.

A Switchgear has complexinternal topology at a networkjunction, so it is derived fromthe ComplexJunctionFeature

class.

These objects are part of thegeodatabase data access objectsand are supplied with ArcInfo.Most data modelers can build richdata models with only thestandard feature and objecttypes.

A programmer develops custom features forspecialized drawing, inspection, validation,

analysis, or messaging.


Geodatabase data

Graph

This is a simplified UML diagram extractedfrom the ArcInfo Object Model diagram thathighlights the key geodatabase data accessobjects. This diagram is most useful forprogrammers, but also gives data modelersinsight into the structure of a geodatabase.

A graph represents a set oftopologically related feature

classes.

A feature dataset iscomposed of graphsand feature classes.

Geometric-Network

Junction-Feature

Network-Feature

EdgeFeature

Complex-Edge-

Feature

Simple-Edge-

Feature



Junction features represent thenodes of a network.

A geometric network is atype of graph that representsa one-dimensional network

such as a utility ortransportation system.

A network feature is a type of feature withspecific properties and behavior relevant to one-dimensional networks. Network features resideonly within feature classes referenced by ageometric network.

Edge features represent thelines that make up a network.

A row is a record in a table. All rows in atable share the same set of fields.

Raster-DatasetTinDataset Feature-

Dataset

GeodatasetA geodataset is a type of dataset that storesgeographic data.

Domain

Coded-Value-

Domain

Range-Domain

*

1..*

Versioned-Workspace

Workspace

Dataset

A workspace in the geodatabase data modelcorresponds to a geodatabase, an ArcInfo coverageworkspace, or a folder with shapefiles.

Row

Feature

Object

Attributed-Relationship

Relationship

*

A feature is a typeof object with ageometric shape.

Relationshipslink featuresand controltheir behaviorwhen they aremoved ordeleted.

Domains are constraints to enforcean expected set or range of attributevalues.

A versioned workspace referencesmultiuser geodatabases for whichversioning is enabled.


1..*

Attributed-Relationship-

Class

Relationship-Class

Feature-Class

2 *

Table

ObjectClass

access objects

1..*

Geometry-Def

0..1

1..*

Fields

Field

Class-Extension0..1

Feature-Class-

Extension

ObjectClass-Extension

Junction-Connectivity-

Rule

Edge-Connectivity-

Rule

Attribute-Rule

Relationship-Rule

Connectivity-Rule

Rule*

Rules are associated with tables or featureclasses. They can be used to validategeographic objects (features) ornongeographic objects (rows).

An attribute rule is theapplication of an

attribute domain to anattribute.

A connectivity ruleconstrains whichnetwork features can beproperly connected withanother.

A relationship ruleconstrains the cardinality

of a relationship.

A table is a collectionof rows, which haveattributes stored incolumns. A featureclass is a table thatincludes a column for ageometric shape.

Each table is associatedwith a set of fields, whichare the descriptions ofhow attributes in columnsare formatted.

If a field representsgeometry, GeometryDefspecifies its particulars.

UML diagram keyis a type of

creates

is associated with

is composed of

1..*

CreateableClass

AbstractClass

InstantiableClass

An abstract class is a specification of the methods andproperties to be inherited by other classes. You cannotcreate objects from an abstract class.

You can directly create objects from a createable classwith a statement like “Dim as New <object>”.

You can indirectly create objects from an instantiableclass by calling a method of another class.

Sample multiplicities: “1” or “ ” is one, “0..1” is zero orone, “ * ” is zero to any integer, “1..* ” is one to any integer.

multiplicity

Implementing classextensions is a wayfor a programmer toadd classwidecustomization, suchas a custom form fordata inspection.

An attributedrelationship class is atype of table that storesrelationships.

An object class is a type of tablethat stores nonspatial objects.

A feature class is a typeof object class that storesspatial objects.

A relationship classrepresents relationshipsthrough embeddedforeign keys.

101

6 The shape offeatures

A GIS enhances database technology byadding spatial data types. The ArcInfogeometry model is the foundation for thedisplay, editing, and analysis of discretefeatures.

These are the topics in this chapter:

• Geometry and features

• Constructing geometry

• Testing spatial relationships

• Applying topological operators

• Geometry object modelA Comparative View of the Principal Waterfalls, Islands, Lakes, Rivers,and Mountains in the Western Hemisphere, John Rapkin, 1851.


GEOMETRY AND FEATURES

One of the primary data representation models ofgeography is the vector data model. In ageodatabase, vector data is implemented as featuresstored in feature datasets and feature classes.

Features offer several important advantages for datamodeling:

• Features are stored as distinct entities withattributes, relationships, and behavior. This letsyou create a rich model that captures all that youknow about a set of geographic features.

• Features have precise locations with well-definedgeometric shapes. In ArcMap, you can applyspatial operators that test for inclusion, overlap, oradjacency among a selection of features.

• Features can be drawn on a map with any color,line width, fill pattern, or other cartographicsymbol. You can create maps that display featureattributes with symbols. You can also print a mapof features with crisp detail at a wide range ofmap scales.

Features are especially well suited for modeling man-made objects. That is because roads, buildings,airports, and other built objects have sharp, well-defined boundaries.

The foundation of representing a feature in ageodatabase is its geometry, or shape. This chaptersurveys the fundamentals of the ArcInfo geometrysystem and its key functions.

This chapter gives you insight in three ways:

• When you build your data model, you willunderstand how to best represent the shape ofyour features.

• When you edit maps in the ArcMap Editor, thesection on geometry construction will help makeyou an expert map editor.

• If you are an application developer and need tocustomize how feature shapes are created andupdated, this chapter gives you an overview ofthe geometry object model. This is not completedocumentation of the geometry system, but it isan overview of the important concepts andsoftware interfaces.

THE GEOMETRY SYSTEM

Each feature has a geometry (or shape) associatedwith it. In the geodatabase, the geometry is stored asa special field in a feature class that is called “shape.”

In the geometry object model, there are two levels ofgeometries—those that define the shape of features,and those that are components of those shapes.

FEATURE GEOMETRIES

A feature can be created with one of these types ofgeometries: point, multipoint, polyline, and polygon.An envelope is a geometry that describes the spatialrange of feature geometries.

An important advantage of the geodatabase datamodel over the coverage data model is that single-part and multipart geometries are combined withinthe same feature class. A feature class with polylinegeometry can contain single-part or multipartpolylines. A feature class with polygon geometry cancontain single-part or multipart polygons. This givesyou more freedom to model the shape of featuresand simplifies the structure of your geodatabase.

Points and multipoints

Points are zero-dimensional geometries. They havean x,y coordinate, with an optional altitude (z),measure (m), and point IDs. Points are used torepresent small features such as wells and surveypoints.

Point Multipoint

Multipoints are unordered collections of points.Multipoint features represent a set of points with acommon set of attributes, such as a set of wells thatform a single unit.

Polylines

A polyline is an ordered collection of paths that canbe disjoint or connected. Polylines are used torepresent the geometry of all linear features.

Chapter 6 • The shape of features • 103

Polyline withone path

Polyline with multiple connected paths

Polyline with multiple disjoint paths

Polylines are used for roads, rivers, and contours.Simple linear features are represented by polylineswith one path. Complex linear features, such asroutes, are represented by polylines with manypaths.

Polygon

A polygon is a collection of rings that are partiallyordered by their containment relationship.

Polygon with one ring

Polygon with interior ring and island ring

Polygon with multipledisjoint rings

Polygons are used to represent the geometry of allareal features. Simple areal features are representedby polygons with one ring.

When rings are nested, they alternate betweeninterior rings and island rings. Rings in a polygoncan be disjoint but they cannot overlap.

Envelope

An envelope represents the spatial extent of features.

Geometry

Envelope

XMin XMax

YMin

YMax

UpperLeft

LowerLeft

UpperRight

LowerRight

An envelope is a rectangle that spans the minimumand maximum coordinates of a geometry. Anenvelope also records the range of z and m valuesfor a geometry. The sides of an envelope are parallelto a coordinate system.

All geometries have envelopes. Envelopes are usedin ArcInfo to enable rapid display and spatialselection of features.

COMPONENTS OF FEATURE GEOMETRIES

Segments, paths, and rings are the geometries thatare components of the feature shapes.

Segments

A segment consists of a start and endpoint and afunction defining a curve between the points.

Line Circular arcBézier curveElliptical arc

The four types of segments are lines, circular arcs,elliptical arcs, and Bézier curves.

• A line is a straight segment bounded by twoendpoints. It is the simplest type of segment.Lines are used for straight constructions, such as ahighway segment, or subdivisions of land, such asa parcel line.

• A circular arc is a portion of a circle. The mostcommon use of circular arcs is for road curbs atstreet intersections. Circular arcs are widely usedin COGO (coordinate geometry) applications.When a circular arc is part of a feature, it is nearlyalways tangent to the connecting segments.

• An elliptical arc is a portion of an ellipse. It is notfrequently used for features, but can approximatetransitional figures such as sections of a highwayramp.

• A Bézier curve is defined by four control points. Itis a parametric curve defined by a set of third-order polynomials and is useful for depictingsmoothly varying features such as contours andstreams. Bézier curves are also used for theplacement of text characters of a name along ameandering object such as a stream.

Paths

A path is a sequence of connected segments. Thesegments in a path cannot intersect. A path cancontain any combination of lines, circular arcs,elliptical arcs, and Bézier curves. Paths make uppolylines.


Path with one�line segment

Path with one�circular arc and two �tangent line segments

Path with two tangent �Bézier curve segments

Often, the segments that comprise a path are tangentto each other. That means that the segments join atthe same angle. For example, a road is typicallycomprised of straight lines and circular arcs. When aline and a circular arc in a road join, they are at thesame angle, or tangent to each other.

Another example of a path with tangent segments isan elevation contour, which is usually a set ofconnected and tangent Bézier curves.

Rings

A ring is a path that is closed and has anunambiguous inside and outside.

The coordinates for the start and endpoints of thepath are the same. Rings make up polygons.

ATTRIBUTES OF FEATURE GEOMETRIES

When you create a feature class, you can assign upto three optional attributes to the vertices of featuregeometries: a z value, an m value, and a point ID.

Vertical measurements with z values

The ArcInfo geometry system is fundamentally atwo-dimensional system, but you can assign a zvalue for each point in a point, multipoint, polyline,or polygon.

Point with z value Polyline with z values Polygon with z values

5280

659

661

664

667

550

550

550550

550

Z values most commonly represent elevations, butthey can also represent another quality, such asrainfall level.

Z values can be applied to features such as streamlines, ridge lines, or lakes. A ridge line is a profilealong a surface. You can assign individual elevationsat each point along a ridge line. A lake polygonwould have identical z values along the perimeter ofthe lake.

One use of z values is to prepare elevation data forinput into a triangulated irregular network (TIN).Another use of z values could be a civil engineeringapplication that models the vertical profile of a roadalignment.

Linear measurements with m values

Some applications employ a linear measurementsystem that is based on interpolated distances alongpaths. You can assign an m value, or linear measure,to each point in a point, multipoint, polyline, orpolygon.

Polyline with m values

2500 2600

2678

28572757

An example of a linear measurement system ismileposting, or stationing, along a road or canal. Thegeometry system has functions to interpolatem values for x,y points along a path or to calculatex,y positions from an m value along a path.

Managing points with point IDs

Sometimes, unique identifiers are collected withpoints. For example, each point collected with asurvey instrument often has a point number. You canuse point IDs to preserve any type of identifier thathas been collected, and point IDs can be used in acustom ArcInfo application.

Note: At the initial release of ArcInfo 8, the ArcMapEditor does not directly edit z values, m values, orpoint IDs. When you edit features, the z values, mvalues, and point IDs that exist for a feature geometryare preserved. If a feature with m values is split, them values of the points of the split features areinterpolated correctly.


CONSTRUCTING GEOMETRY

One of the important services of the ArcInfogeometry system is a rich set of constructionmethods that create new geometries from distances,angles, and relationships to existing geometries.

Many of these methods correspond to basiccoordinate geometry (COGO) commands, which areuseful for accurately converting survey data togeographic data. Other methods facilitate the editingof features to improve cartographic presentation orto control spatial relationships between features.

ArcInfo users will see most of these constructionmethods implemented as tasks within the ArcMapEditor. This section will give you insight into howthe edit sketch in the ArcMap Editor works with thegeometry system.

Programmers can build custom applications thatmodify the shapes of features by using theconstruction methods in the geometry object model.This section is not a complete reference to thegeometry construction methods, but does list someof the most commonly used constructors.

Units and input

deflection�angle

baseline

geometric �angle

point

Angles are normally specified as geometric anglesmeasured counterclockwise from the positive x axisof the Cartesian coordinate system. Someconstructors use deflection angles that are measuredfrom a point relative to a baseline.

If you are a programmer, all angles are specified inradians. If you are an ArcInfo user, angles arespecified in degrees by default.

All distances are measured in map projection units.

Some constructors are built upon segments, paths,and rings. Recall their definitions: a segment is a line,circular arc, elliptical arc, or Bézier curve; a path is asequence of connected segments; and a ring is aclosed path.

Point construction

The methods that follow create a single point byspecifying angles, distances, and relationships toexisting geometries.

Programmers can access point constructors throughthe IConstructPoint interface implemented in thePoint class.

Construct Along

distance �(or ratio)


curve

curve

Given a curve and a distance or ratio, a point isconstructed along that curve. If the distance isgreater than the length of the curve, the pointfollows either a tangent or the embedding geometry.

Construct Angle Bisector

1/2α

1/2α

from-pointthrough-point

to-point

length

Given a from-point, through-point, and to-point, thisconstructor bisects the angle subtended by the threepoints and places a point along the bisector at thelength specified. A negative length places the pointalong an obtuse angle bisector.

Construct Angle Intersection

βα

Given two points and two angles, this constructorplaces a point at the intersection of the rays definedby the points and angles.


Construct Deflection

αdistance

baseline

Given a line that serves as a baseline, a deflectionangle, and a distance, this constructor places a pointat the distance along a ray at the deflection angle.

Construct Deflection Intersection

Given a line that serves as a baseline, a deflectionangle measured from the start point of the baseline,and a deflection angle measured from the endpointof the baseline, this constructor places a point at theintersection of the defined rays.

Construct Offset


offset

path


offset

path

Given a path, a distance or ratio along the curve,and an offset distance, this constructor places a pointat the offset. A positive value constructs a right offsetand a negative value constructs a left offset. In civilengineering, this constructor is called a station offset.

Construct Parallel

distance (or ratio)

point

path

Given a path with straight lines, a reference point,and a distance, a new point is placed along aparallel curve. Note that the offset distance does not

need to be specified; it is inferred from the geometryof the reference point and curve.

A similar polyline constructor, Contruct Offset, cancreate a parallel figure with paths containing circulararcs as well.

Multipoint construction

The constructors that follow return a set of points asa multipoint geometry.

Features with simple point geometries are far morecommon than features with multipoint geometries.When you apply these constructors, your applicationor tool may likely select one point from a multipoint,and then create or modify a simple point feature.

These constructors are present in theIConstructMultipoint interface implemented in theMultipoint class.

Construct Circular Arc Points

Point of �intersection

Radius�point

Start �point

Endpoint

Given a circular arc, this constructor returns theendpoint, start point, radius point, and point ofintersection. (Civil engineers refer to the start pointas the point of curvature and the endpoint as thepoint of tangency.)

Construct Divide Equal

n = 4

Given a curve and an integer number, thisconstructor places that number of points evenlyspaced along the curve.


Construct Divide Length

1*distance2*distance



Given a curve and a length, this constructor placesas many points as possible spaced by that length.

Construct Implied Intersection

Given two segments, this constructor places points atthe actual and extended intersections of thosesegments. A line segment is extended to an infiniteray, a circular arc is extended either tangentially orusing its embedding circle, an elliptical arc isextended either tangentially or using its embeddingellipse, and a Bézier curve is extended tangentially.

Construct Intersection

Given two segments, this constructor places pointsonly at the actual intersections of those segments.

Construct Tangent

Given a circular arc and a point, this constructorplaces points at the positions of tangencies from thepoint to the arc.

Construct Three Point Resection

βα

Given three points and two angles measured from astation point at unknown location, calculate andplace the station point.

Line construction

A line is a straight segment between two points.Lines are building blocks for polylines and polygons.

This constructor is present in the IConstructLineinterface implemented in the Line class.

Line ConstructAngleBisector

1/2α

1/2α

from-pointthrough-point

to-point

length

Given a from-point, through-point, and to-point andlength, bisect that angle and construct a line at thatlength.

Circular arc construction

Circular arcs are segments that are part of theboundary of a circle. Like lines, circular arcs arebuilding blocks for polylines and polygons.

Construct Arc Distance

start point

center point

arc distance

Given a center point, start point, and arc distance,this constructor builds a circular arc in acounterclockwise direction. The arc distance must begreater than zero. A Boolean value specifies whetherthe arc is constructed clockwise or counterclockwise.


Construct Chord Distance

start point

center point

chorddistance

Given a center point, start point, and chord distance,this constructor builds a circular arc in a clockwiseor counterclockwise direction, as specified by aBoolean value.

A chord is an imaginary line segment between thestart point and endpoint of a circular arc.

Construct Chord Height

chord height

from-pointto-point

Given a from-point, to-point, and a height above themidpoint of a chord, this constructor builds a circulararc in the clockwise or counterclockwise direction,as specified by a Boolean value.

Construct Fillet

segment 1segment 2

radius

Given two segments and a radius length, thisconstructor builds a circular arc tangent to the twosegments.

A fillet is a circular arc that is tangent to twosegments. Most often, the tangent segments are lines,but they can also be circular arcs.

Construct Tangent and Point

point

segment

Given a segment and a point and a Boolean valuespecifying the start or endpoint of the segment, thisconstructor builds a circular arc.

Construct Three Points

start point

middle point

endpoint

Given a start point, middle point, and endpoint, thisconstructor builds a circular arc that uniquely fitsthose points.

Construct Two Points and Radius

start point

radius

endpoint

Given a start point, endpoint, and radius length, thisconstructor builds a circular arc. A Boolean valuespecifies whether the center point on the arc is onthe left or right of the chord from the start point tothe endpoint.

Curve construction

A curve is either a simple segment, a path, theboundary of a ring, a polyline, or the boundary of apolygon.

This constructor is present in the IConstructCurveinterface that is implemented in the Polyline andPolygon classes.

Construct Offset

offset

input curve

output curve

input curve

output curve

offset

miter bevel round

Given an input curve, this constructor constructs anoffset figure that is offset by the specified distance.

You can specify whether the offset curve is mitered,beveled, or rounded. You can specify a beveling orrounding distance.


Path construction

A path is a sequence of connected segments.

This constructor is present in the IConstructPathinterface that is implemented in the Path class.

Construct Rigid Stretch

stretch vector at vertex stretch vector inside segment

stretch vector at end node stretch vector inside end segment

Given a path and a stretch vector, this creates a paththat is proportionally stretched as shown.

This constructor is very useful for interactive rubbersheeting and conflation, which are editingtechniques that adjust existing geometries to betterconform to positions with higher accuracy, such assurvey data.

Angle construction

These constructors return an angle and are presentin the IConstructAngle interface implemented in theGeometryEnvironment class.

Construct Line

�line

α

Given a line, calculate its angle.

Angle Construct Three Point

from-point

through-point

to-point α

Given a from-point, through-point, and to-point,calculate the sweep angle between the points.


TESTING SPATIAL RELATIONSHIPS

The ArcInfo geometry system includes a set ofBoolean operators that test the spatial relationshipsbetween a base geometry and a comparisongeometry. These operators can be applied to points,multipoints, polylines, and polygons.

The base geometry is the object invoking theoperator. The comparison geometry is the geometryexpressed as a parameter in the operator. The resultof the relational operator is returned as a Booleanvalue. No new geometries are created with theseoperators.

Equals

Does the base equal the comparison geometry?

Base Geometry

Com

paris

on G

eom

etry

No equals�relationship�

possible


possible


possible


possible


possible


possible

For the base geometry and comparison geometry tobe equal, all of their constituent points must haveidentical coordinate values.

The geometries that are compared must have thesame dimension.

Contains

Does the base geometry contain the comparisongeometry?

Base Geometry

Com

paris

on G

eom

etry

No containment�relationship�

possible


possible


possible

For the base geometry to contain the comparisongeometry, it must be a superset of that geometry.

A geometry cannot contain another geometry ofhigher dimension.

Within

Is the base within the comparison geometry?

Base Geometry

Com

paris

on G

eom

etry

No within�relationship�

possible


possible


possible

For the base geometry to be within the comparisongeometry, it must be a subset of that geometry.

A geometry cannot be within another geometry oflower dimension.


Crosses

Does the base geometry cross the comparisongeometry?

Base Geometry

Com

paris

on G

eom

etry

No crossing�relationship�

possible


possible


possible


possible


possible


possible

For the base geometry to cross a comparisongeometry, they must intersect in a geometry of lesserdimension than the highest dimension.

Two lines can intersect at points. A line and an areacan intersect at lines.

There is no crossing relationship possible between abase area and comparison area. This is consideredan overlap relationship.

Disjoint

Is the base geometry disjoint from the comparisongeometry?

Base Geometry

Com

paris

on G

eom

etry

A base geometry is disjoint from a comparisongeometry if they share no points.

Overlaps

Does the base geometry overlap the comparisongeometry?

Base Geometry

Com

paris

on G

eom

etry

No overlap�relationship�

possible


possible


possible


possible


possible


possible

A base geometry overlaps a comparison geometry iftheir intersection is a geometry of the samedimension. An overlap relationship requires thatboth geometries be of the same dimension.

Touches

Does the base geometry touch the comparisongeometry?

Base Geometry

Com

paris

on G

eom

etry

No touch�relationship�

possible

Two geometries touch when only their boundariesintersect.


APPLYING TOPOLOGICAL OPERATORS

The geometry system provides a set of operators thatreturn geometries based on logical comparisonsbetween sets of points in one or more geometries.

These operators provide support for editinggeographic features that overlap. They are present inthe ITopologicalOperator interface, which isimplemented in the Envelope, Multipoint, Point,Polygon, and Polyline classes. In the GIS literature,these are sometimes called spatial topologicaloperators.

Buffer

d

Given a geometry and a buffer distance, the bufferoperator returns a polygon that covers all pointswhose distance from the geometry is less than orequal to the buffer distance.

Clip

Given an input geometry and an envelope, the Clipoperator returns a new geometry with the set ofpoints of the input geometry that are within or onthe boundary of the envelope.

Convex hull

Given an input geometry, the convex hull operatorreturns a geometry that represents all points that arewithin all lines between all points in the inputgeometry.

A convex hull is the smallest polygon that wrapsanother geometry without any concave areas.

Cut

LeftRight

Left

Right

Given a cut curve and a geometry, the cut operatorwill split the geometry into a right part and left part,relative to the direction of the cut curve.

Points or multipoints cannot be split. Polylines andpolygons must intersect the cut curve to be split.

Only two geometries are created by the cut operator,but they can have multiple parts.


Difference

Base

Comparison Result

The difference operator returns a geometry thatcontains points that are in the base geometry andsubtracts points that are in the comparison geometry.

Intersect

The intersect operator compares a base geometry(the object from which the operator is called) withanother geometry of the same dimension and returnsa geometry that contains the points that are in boththe base geometry and the comparison geometry.

Symmetric difference

The symmetric difference operator compares a basegeometry (the object from which the operator iscalled) with another geometry of the samedimension and returns a geometry that contains thepoints that are in the base geometry or the points inthe comparison geometry, but excludes the points inboth geometries.

Union

The union difference operator compares a basegeometry (the object from which the operator iscalled) with another geometry of the samedimension and returns a geometry that contains thepoints that are in the base geometry together withthe points in the comparison geometry.


Geometry object modelThis is a simplified UML diagram extractedfrom the ArcInfo Object Model diagram thathighlights the key geometry objects. Thisdiagram is most useful for programmers, butalso gives data modelers insight into thestructure of feature shapes.

*Segment

Envelope Point

Ring

Curve

Polyline PolygonEllipticArcCircularArc

Geometry

*

Line BezierCurve

Path

*

Multipoint

Polycurve

*

A multipoint is a collectionof points.

Bézier curves, circular arcs, elliptical arcs, and lines are types ofsegments that, when joined at endpoints, form a path.

A ring is a simplenonintersecting area.

A point is a discretex,y location.

An envelope is arectangle that spans afeature.

A feature is associated with oneof these types of geometry:points, multipoints, polylines,and polygons.

A polyline is an arbitrarycollection of paths.

A polygon is an arbitrarycollection of rings.

A path is a set ofsegments that areconnected.


creates

is associated with

is composed of

1..*

AbstractClass An abstract class is a specification of the methods and

properties to be inherited by other classes. You cannotcreate objects from an abstract class.

You can directly create objects from a createable classwith a statement like “Dim as New <object>”.

You can indirectly create objects from an instantiableclass by calling a method of another class.

Sample multiplicities: “1” or " " is one, “0..1” is zero orone, “ * ” is zero to any integer, “1..* ” is one to any integer.multiplicity

CreateableClass

InstantiableClass

115

7 Managingwork flowwith versions

GIS is being applied by organizations of everysize. These organizations apply internalbusiness practices on units of work thatrepresent projects, work orders, designs, orstudies. Geodatabases can be versioned sothat many people can concurrently editgeographic data and manage work flows.


• Using versions

• Long transactions and the geodatabase

• The fundamentals of versions

• Editing versioned geodatabases

• Types of work flowsTopographic Representation of the New Russian Capitol and SeaportSt. Petersburg, Johann Baptist Homann, before 1724.


USING VERSIONS

Many applications of GIS involve long-term designefforts that require the cooperation of a number ofpersons and departments. These design activitiestake place at the organizations that build things—utilities, municipal and regional governments, anddepartments of transportation.

These organizations have established work-flowprocesses for design, construction, and maintenance.The general steps include the initial engineeringdesign, exploration of design alternatives, selectionand approval of a design, the construction of thedesign, and updating maps with the constructionfeatures as they have been built in the field.

When you use GIS in these work-flow scenarios, it isnecessary that multiple persons be able tosimultaneously edit a geodatabase. They also needto have a transacted view of the geodatabase so thatonly the changes that they or their coworkers makeare visible to them. Further, the work-flow structureneeds to emulate the business practices of variousdepartments in an organization.

The geodatabase data model serves these needsthrough a data management framework calledversioning. This framework lets you create versionsof a geodatabase for the states of a project, reconciledifferences between versions, and update the masterversion of a geodatabase with the design as built.

This chapter documents the fundamentals ofversioned geodatabases and shows how they can beemployed with some work-flow scenarios.

DESIGN SCENARIO

To illustrate how versioned geodatabases are used ina multiuser environment, follow the scenario from awater utility shown on the facing page.

A municipal water utility keeps a comprehensivegeodatabase with the current state of its field assets.All of the water pipes, valves, pumps, and othercomponents of the water system are recorded asfeatures in a geodatabase that is updated daily.

This water utility has a number of departments thatare responsible for different phases of constructingand maintaining the water system. Because of thisorganizational structure, this utility uses a versionedgeodatabase served through ArcSDE.

A versioned geodatabase has a top-level version thatis always called “default.” The default version of thegeodatabase represents the water system in its bestknown as-built state. It is the starting point forcreating new designs and construction activities.

Continuous editing of the geodatabase

The mapping department, Liz and Maria, areresponsible for the daily maintenance of thegeodatabase. To support the new line extension,Maria reviews field notes from that area and updatesthe water meter features. Liz adds new survey datapoints that were collected in a field survey inadvance of the line extension. These edits are madedirectly to the default version of the geodatabasebecause they represent improved knowledge of thewater system and are not part of a design cycle.

Creating versions by department

The information systems department is responsiblefor the corporate database that supports customerbilling and asset management. To support the lineextension, Fritz creates a new version, uses locatorsto match billing records to network features, andextracts and summarizes water usage data. Thissummarized data is intended only for this lineextension project, so this version is temporary anddiscarded at the end of the design process.

The engineering department takes the data collectedby the other departments and creates two versionsfor two engineering designs. Petra and Taylor createan engineering design based on using 16-inch pipefor the new main line. They simultaneously work onthis version and create a proposed design. Fillycreates another engineering design based on a24-inch pipe to examine whether the increased pipecost is offset by greater efficiency in handlingpresent and future water usage. She discovers thatthe 24-inch pipe will serve projected water demandfor 12 more years and that the greater initialconstruction cost is justified. Her design gets postedto the line extension version. When construction iscomplete, the line extension project version is postedinto the default version.

This scenario is a simple example of how versionscan be used to support a rich modeling environmentfor organizations that build complex systems.

Chapter 7 • Managing work flow with versions • 117

Versioning scenario

The municipal water utility is planning an upgrade of a part of its system. Several departments areparticipating in data preparation and the design of two alternate line extensions. This scenario illustrates

how a multilevel version tree lets many users collaborate on solving an engineering problem.

updates of data. The updates getsaved to the default version atthe end of every edit session.

Default

1998 water demand

5530

2320

1475

1200

8900

6780

4350

ArcMap session - Liz

addsurveydata

updatemeter

records

ArcMap session - Maria

ArcMap session - FillyArcMap session - Petra ArcMap session - Taylor

Line extension project

16-inch pipe design 24-inch pipe design

ArcMap session - Fritz

H2O usage

1The default version is the nominal state

of the geodatabase. It represents thecurrent status of the entire water utility

as best it is known.

2 3

5

11

8 9

4

6

12

7

10

Meanwhile, the information systems groupcreates a version to analyze water demandfrom historic records. After the analysis iscomplete, this version is discarded.

Fritz extracts peak water usagefrom city databases and useslocators to match historic waterusage to buildings by address.

Maria reviews engineering datafor water meters and verifies theyare correct and up-to-date for theline extension.

Liz adds survey data that wascollected to support the lineextension project.

Filly creates an alternate design using a 24-inchpipe and discovers new engineering efficiencies

for serving water customers.

Taylor studies the gravity model andexamines the profile of the ground and

the 16-inch water line design.

Petra develops a design using a 16-inchpipe. She explores whether the pipesizes can handle the peak demand.

Two versions are created from the lineextension version to study designs

based on a 16-inch or 24-inch watermain pipe diameter.

The 24-inch pipe design wins approvaland gets posted back to the lineextension version. The 16-inch designis then discarded.

The city engineer creates a newversion for the water lineextension project. On this

version, engineering and surveydata is added as a base for the

design.After the 24-inch lineextension is built, theline extensionversion gets postedback to the defaultversion, with edits forany changes made inthe field.

Maria and Liz make maintenance


LONG TRANSACTIONS AND THE GEODATABASE

ArcInfo 8 is a milestone in the integration of GIS andrelational database technology. GIS has now joinedthe mainstream of information technologies.

These are some of the benefits of storing all of yourgeographic data in commercial relational databases:

• You can integrate geographic data with corporateor agency databases.

• You can use standard database administrationtools for managing your geographic data.

• You can create very large geographic databasesthat can be displayed and edited quickly.

• You can deploy geodatabases on the commercialrelational database of your choice.

• You can serve geographic data to a wide varietyof clients, such as view-only applications, CADapplications, or Internet applications.

The geodatabase extends standard relational (andobject-relational) databases to support the specialrequirements of representing geographic data. Theseare some of the capabilities that a geodatabase addsto a relational database:

• You can represent and store geographic data inthe form of raster datasets, feature datasets, TINdatasets, and location data.

• You can execute spatial and topological analysison geographic data.

• You can perform rich cartographic display andproduce high-quality maps.

• You can add intelligence to features by definingattributes, topological associations, relationships,and validation rules.

• You can enable many users to simultaneouslydisplay and edit geographic data.

This last capability, providing multiple users withread and write access to a geodatabase, is calledversioning and is a critical requirement for manyorganizations. Versioning is a key function ofgeodatabases served through ArcSDE.

This chapter discusses how versioned geodatabaseswork with examples of work-flow scenarios.

TRANSACTIONS IN A DATABASE

A central idea of relational databases is atransaction. Simply put, a transaction is a group ofatomic data operations that comprise a completeoperational task.

Transactions preserve the consistency and integrityof the database by ensuring that either all or none ofthe atomic operations are executed for a task.

Short transactions

When you access data in a database, you have twobasic goals: that the data be accurate and that it istimely.

Relational databases satisfy that requirement withshort transactions, which represent operational tasksthat can be completed in fractions of a second, or aminute or two at most. During the very brief timethat a short transaction is being committed, no otherupdates to the affected rows are possible.

Short transactionscash machine transactions

stock market exchange

updating personnel recordsaccess authorization

hospital records

time measured in seconds

Short transactions represent most of the informationtasks that people engage in, such as drawing moneyfrom an account at an automated teller machine,updating hours worked in a payroll application, orentering medical records.

Once a short transaction is committed in a relationaldatabase, it is not easy to undo that transaction or toreconstruct the state of the database at a historicpoint in time. There is only one state with arelational database: its status as of the most recentlycompleted short transaction.

The short transaction model works very well formany critical applications that require instant accessto a uniform view of data, but geographic datarequires a longer view of updating data.


Editing geographic data

While a short transaction is in progress, the relationaldatabase applies locks on affected rows in databasetables so that data being updated is protected fromchanges until the transaction is complete. When theshort transaction is completed, the locks arereleased.

When multiple people are simultaneously editinggeographic data, this type of row locking isimpractical because even for short edit tasks, thelocks must be held for several minutes.

Another reason that row locking is deficient for GISis that features in a geodatabase coexist in a richcontext of network connections, topologicalassociations, and relationships.

To understand why, consider this scenario: You areediting an electric utility and adding lines, poles,transformers, and other devices. If another personwere to edit a nearby feature while you were editingthis transformer, your network could quickly fall intoan inconsistent state.

In a network, the integrity of onefeature is dependent on others.When you edit a transformer, youmust be certain that no one else ischanging the pole it is mounted on,the voltage level of the circuit, andthe line phasing of connected lines.

Short transactions fail for this scenario because whenyou are adding this transformer, you are not reallyediting a single feature, but you are editing a larger,more complex object—the network as a whole.

Another reason that short transactions are deficientin the multiuser geographic editing environment isthat you must be able to always see the current stateof the database as displayed on a map. Every timesomeone else made a change, your system wouldhave to redraw the map, which may take a numberof seconds for a complex map. This is unacceptable.

Long transactions

What you need for multiple users to edit geographicdata is a transaction type that can do the following:

• Allows multiple persons to simultaneously edit thesame complex system such as a network.

• Spans all of the edits that you need to perform ona work unit, whether it takes an hour or a month.

• Lets you have a private view of your data so thatno one else sees incomplete work.

• Permits you to define the scope of work to matchyour business’ work order system.

This type of transaction is a long transaction.

Long transactions

time measured in hours, days, or months

engineering analysis

construction design scenarios

demographic projections marketing studies

environmental restoration

Long transactions have other uses besides representingconstruction work units. You can use long transactionsto model any type of “what if” scenario.

During the scope of the long transaction, you canfreely add proposed features, perform geographicanalysis, and produce maps—all without affectingyour nominal database. When the scenario is done,you can post the changes to the database if it is builtor discard it if it is not.

Concurrency model

Long transactions implement a data managementapproach called optimistic concurrency. This meansthat when you start a long transaction, no locks areapplied to features. The absence of locks permits theintroduction of editing conflicts, but this is mitigatedby an environment to easily detect, reconcile, andpost these conflicts.

Optimistic concurrency is suitable for GIS applicationsbecause the volume of edits is small compared to thesize of the geographic database. In real work-flowpractices, edit conflicts are not frequent, and the costof reconciling conflicts is minor when compared tothe savings from not having to lock or check outfeatures for the duration of a long transaction.


THE FUNDAMENTALS OF VERSIONS

Versioning is ArcInfo software’s implementation oflong transactions against central multiuser relationaldatabases served by ArcSDE. It is an advanced datamanagement system that lets you adopt any of avariety of work-flow practices when editinggeodatabases in a multiuser environment.

Versioning can be implemented on multiusergeodatabases served through ArcSDE. You cannotimplement versioning on personal geodatabases.

BASIC CONCEPTS

Versioning lets multiple users directly edit ageodatabase without explicitly applying feature locksor duplicating data. The following are the essentialfacts about versions.

A version is a named state of a geodatabase

Proposed malldesign

McPheeterEstates

Work Order19706

Wildlife afterNambe Dam

Snapshot:July 16, 1997

You can use versions to represent engineeringdesigns, construction jobs, snapshots in time ofgeodatabases, and any type of scenario that involvesthe posing of “what if” questions in studying a result.

A version spans a geodatabase and has properties

Version

Geodatabase

Feature dataset

Feature class

Feature class

Feature class

Table

Properties of a versionnameownerdescriptioncreation datelast-modified dateparent versionversion’s permission

privateprotected

public

only the owner can view and editall users can view, but only the owner can editall users can view and edit

permission settings

In ArcCatalog, you can define which objects in ageodatabase are versioned. You can selectivelyspecify which feature datasets, feature classes, andtables are versioned.

When you specify that a feature dataset is versioned,all of its tables and feature classes are automaticallyversioned.

You can control the visibility of a version to otherusers by setting its permission.

A geodatabase can have multiple coexistingversions

Approveddesign

As builtProposal

Underconstruction

Geodatabase

Feature dataset

Feature class

Feature class

Feature class

Table

Each version lets you perform all of the same displayand analytic functions as a nonversionedgeodatabase.

Versions differ from each other only in row state,not in schema

Version

Geodatabase

AddedModified

Removed

A version is aview of a

geodatabasewith changes

applied.

A version presents you with a seamless view of allthe edits applied since the version was created. Therow state reflects all added, removed, and modifiedobjects. The row state information about each


version is stored (or persisted) in the geodatabase.The schema, the definition of tables and their fields,can be modified on a geodatabase; schema changesare applied to all versions of the geodatabase.

Internally, the geodatabase has tables to keep trackof modified, added, and removed features for eachversion, but that is not apparent to you when youuse versions. It appears to be an integral copy of ageodatabase.

Every versioned geodatabase has a default version

Default

Geodatabase

Feature dataset

Feature class

Feature class

The default version can be thought of as the “as-built” version. It usually represents the nominal stateof the geodatabase.

The default version is the geodatabase. Most userswill edit the default version.

A version is created from another version

Alternateproject

Default

Project design

Projectapproval

Starting with the default version, you can create anynumber of versions. Every version, except for thedefault version, has exactly one parent version. Youcan create a complex hierarchical version tree asappropriate for your organization’s work-flowrequirements.

Versions can be removed, but only if their childversions are removed first. Before removing aversion, the changes can be reconciled and postedto another version or discarded.

A versioned geodatabase can be compressed. Overtime, rows are added to various internal tables thatmanage versions in a geodatabase and many aresuperseded by other rows. These extra rows can beeliminated to conserve disk space and preserve dataaccess performance. This is a task for the ArcSDEadministrator.

A user can connect to any version

Underconstruction

Default

Approveddesign

Geodatabase

Feature dataset

Feature class

Feature class

Feature class

Table

A user will start editing a version based on theproject or project stage they are working on. A usercan work on any version they have been grantedpermission to.

WHY VERSIONS PERFORM WELL

When you start using versioning, you will noticeconsiderable performance improvement and greaterease of use over previous data management systems,such as checked out datasets, tiled libraries ofdatasets, or copied datasets.

The reason that versioning works quickly and well isthat versions do not require any duplication orreplication of data. Internally, a versionedgeodatabase uses internal identifiers and managesadditional tables that record which features andobjects are added, removed, or modified.

It is not necessary for a data modeler, or even aprogrammer or database administrator, to be awareof any of the implementation details of versionedgeodatabases. ArcCatalog, ArcMap, and ArcSDEcollectively provide you with an easy-to-use andcomprehensive user interface to versioning.


EDITING VERSIONED GEODATABASES

The ArcMap Editor is the arena in which youperform operations on versions in accordance withyour organization’s work-flow practices.

The basic versioning operations you can do inArcMap are editing a version, reconciling your editsession against another version and resolving anyconflicts that arise, and then posting the changes inyour edit session to a version.

Controlling a version tree

Property Conflict Edit PreEdit

1Edit

3Post

2ReconcileMost features will

pass throughreconciliation.

Some will haveconflicts that need

to be resolved.You begin byselecting aversion andediting it.

From the editor,you choose a

version to mergewith and initiatereconciliation.

Once your edits havebeen reconciled, youcan choose to post all

edits to the targetversion.

Versionbeingedited

TargetVersion

ArcMap

Editing a version

When you are using ArcMap and start editing, if yourmap references one version, then that version isautomatically opened for editing. If your mapreferences multiple versions, you can choose fromone of these to start editing.

Reconciling versions

Reconciliation is the process of merging features andobjects from a target version into the current editsession. A target version can be any version in thedirect ancestry of the version being edited.Reconciliation must be done before posting changesto another version.

When you reconcile, you should have fullpermission to modify the feature classes in the targetversion that you have been editing in the edit

session. If you do not, you will receive an errormessage and will not be able to complete thereconciliation.

ArcMap

parent

child

parent

child

edit session

When you reconcile theedit session, you can

select any of the directancestor versions as thetarget version to merge

features from.

Potential targetversion

Potential targetversion

Downtownrenovation

Newhospital

Defaultversion

Because a version spans all the versioned featuredatasets, feature classes, and tables in a geodatabase,all objects and features in these classes will bemerged into the edit session. The great majority offeatures and objects will pass straight throughreconciliation from the target version to your editsession.

Handling conflicts during reconciliation

A small percentage of features and objects will haveconflicts when compared between the target versionand the edit session.

There are two types of conflicts:

1 When the same feature is updated in both thetarget version and the edit session.

2 When the same feature is updated in one versionand deleted in the other.

For most reconcile operations, no conflicts will beencountered. That is because at most organizations,projects and versions represent distinct geographicareas. If you and your coworkers are editingdifferent parts of the map, it is generally not possibleto introduce conflicts. Conflicts usually arise when


multiple people are editing features that are in closeproximity.

When conflicts do arise, you will see an interactiveconflict resolution dialog. This dialog lets youexamine and zoom to any conflicts between the twoversions. The conflicts are grouped by conflictclasses, which are feature classes and tables thatconflicts are detected for.

ArcMap session

Version being edited

Target version

Property Conflict Edit PreEdit

FeatureIDShapeSymbolArea...others

The target version is either theparent or direct ancestor of the

version being edited.

The target version is either theparent or direct ancestor of the

version being edited.

Replace with conflict versionReplace with edit version

Replace with pre-edit version

Use feature from target version.Ignore feature from target version.Return feature to original state beforeediting began.

Options for reconciling individual conflicts

For each conflict, you can choose whether to replacethe feature in your edit session with the conflictfeature from the target version, keep it as it is in youredit session, or revert it to its state at the beginningof your edit session.

When you work with features that have relationshipswith features in other feature classes, there areadditional considerations for deciding how to resolveconflicts. For details, read chapter 11, “Working witha versioned geodatabase,” in Building aGeodatabase.

Posting versions

You can post a version to the target version after youhave successfully completed the reconciliation.

ArcMap

parent

child

edit session

When you post the edit session, allof the reconciled changes to the edit

session are saved to the targetversion. The edit session and target

version are now identical.

Target versionDowntownrenovation

Newhospitalversion

The post operation synchronizes the row state ofyour edit session with the target version. They areidentical at this point.

At this point, you may continue to make more editsin your edit session, but you will need to undergothe reconciliation, conflict resolution, and postingprocess again if you want to apply these changes tothe target version.

If a posting marks the end of your project, you cango to ArcCatalog to terminate that part of your workflow by removing the version you have been editing.


When you apply versioned geodatabases to yourorganization, you can select from one or severaltypes of work flows that match your businesspractices.

The following are a summary of the basic workflows supported by versioning. Your implementationcan be one or a combination of these work flows.

Direct editing

The simplest work flow for multiuser access on ageodatabase is for many users to directly edit thedefault version.

As each person opens the default version for editing,a temporary version is created. The editor is notexplicitly aware that a version is created and doesnot give it a name. Whenever the editor saves thework or ends the edit session, then that temporaryversion is automatically reconciled with and postedto the default version.

If there are conflicts, you must resolve them with theconflict resolution dialog before you can successfullysave your edits. If no conflicts are detected, the editsare directly posted to the default version.

This work flow has the virtue of simplicity. It is mostappropriate where the units of work are fairlymodest in scale and where no design alternativeshave been explored or historical snapshots made.

Two-level tree

Many organizations employ a more structuredprocess that tracks discrete work units ofconstruction or maintenance.

These work units typically span a time interval ofdays, weeks, or months and represent tasks such asadding new phone service, adding a new lineextension with pipes or poles and wires, andbuilding a new pump station or electric substation.

When a work order or project is initiated, a versionis created. One or several people work on thisversion until the design or construction is complete.At that point, reconciliation and posting are done tomerge the work order features into the defaultversions, and then the work order version can beremoved.

Multilevel tree

Some organizations’ projects have a higher level ofstructure and can be subdivided into functional orgeographic parts.

For example, a project to design and construct anew shopping mall might have phases ofconstruction, be subdivided into east and west parts,or be subdivided by construction activities such asstructure, gas and water, and electric.

For larger projects with departments and teams, amultilevel version tree is an effective way to organizework flow. The teams that are working on eachaspect of the project have their own version, withwhich they can maintain a private view of theirdesigns and then post the designs when constructed.

Cyclical

Many projects go through a prescribed or regulated setof stages that require engineering, administrative, orlegal approval before proceeding to the next stage.

A version represents each stage of this process. Acyclical work flow can capture the design at eachstage, and when the last stage is reached andfinished, the design can be posted directly to thedefault version, which represents the nominal stateof the database.

This work flow saves the effort of progressivelyposting changes up the version tree; you can bypassthe immediate parent versions and post directly tothe default or other version.

Extended history

For some projects, it is desirable to preserve aversion that reflects a historic state of a project.

You can define a historic version on a projectversion, and when the project version is posted to itsparent version, the historic version remains as asnapshot in time.

SUMMARY

In practice, you will probably either apply the directediting work flow or some combination of theothers. An understanding of the elements of work-flow management will improve the effectiveness ofyour geodatabase design.

TYPES OF WORK FLOWS


Work flows with a versioned geodatabase

Direct editing

Many people connect to and directly edit the default version.When the edit session is saved or ended, all conflicts are

identified and the features are reconciled.

Defaultversion

Two-level tree

The default version represents a master geodatabase.There are many child versions that represent designsor projects. Features can be reconciled in the childversions and optionally posted to the default version.

Defaultversion

Project 1 Project 2 Project 3

Extended historyWith extended history, you can declare historic

versions that need not be posted before theparent version is posted to its parent. A common

use of extended history is to preservean archive of the default version.

ProjectHistoric -Phase 1

ProjectHistoric -Phase 2

Defaultversion

Default as ofDecember 1

Project 2Default as ofJanuary 30

Multilevel tree

Versions can be organized in a multilevel tree.Project 1 cannot be posted to the default

version until parts 1 and 2 are posted.

Project 1Part 1

Project 1Part 2

Defaultversion

Project 1 Project 2 Project 3

CyclicalIn a cyclical work flow, versions followone another. When the constructed version isposted, the intermediate versions can be removed ifthe constructed version is first removed, orpreserved for historical representation.

Proposed

Approved

Defaultversion

Constructed

127

8 Linearmodeling withnetworks

Networks model the movement of people andthe flow of resources such as electricity, gas,water, and communication. ArcInfo representsthese linear systems as geometric networks.


• Modeling infrastructure

• The network model

• How features connect

• Network features

• Network flow

• Analysis on a network

• Network object modelBononiensis Ditio, The Gallery of Maps in the Vatican, 1636.


MODELING INFRASTRUCTURE

The economic foundation of our world is itsinfrastructure: the collection of highways, cables,and pipelines that enables the movement of people,energy, commodities, and ideas.

This infrastructure is modeled as networks, and theform, capacity, and efficiency of these networkshave a substantial impact on our standard of livingand our perception of the world around us.

ArcInfo 8 introduces a new network model—thegeometric network—that builds on years ofexperience modeling transportation and utilitynetworks. Geometric networks let you reach a newlevel of sophistication for naturally modelinginfrastructure.

These are the principal benefits of the geometricnetwork model:

• Editing networks is easy. When you add networkfeatures, you can ensure that they are properlyconnected with network connectivity rules.

• Network features can represent complex parts ofa network, such as switches. This simplifiesediting and lets you create better maps withfewer features in your network representation.

• A suite of simple and advanced network analysissolvers is built into ArcInfo, ready to use.Network analysis is fast even on very largedatasets.

• Networks can be versioned. Many people cansimultaneously edit the same large network incompliance with their organization’s work-flowpractices.

This chapter documents the important qualities ofnetworks and reveals how the geometric networkand the underlying logical network form the basisfor advanced modeling of transportation, energy,and communication systems.

NETWORKS AND APPLICATIONS

Networks are simple. They are comprised of twofundamental components, edges and junctions.

junction

edge

network

Some examples of edges are streets, transmissionlines, pipe, and stream reaches.

Some examples of junctions are street intersections,fuses, switches, service taps, and the confluence ofstream reaches.

Edges connect at junctions and the flow from oneedge—automobiles, electrons, water—can betransferred to another edge.

From this simple idea, you can build networks toserve any of a myriad of applications. Here are afew examples:

• A railroad schedules its trains to efficiently linkwith intermodal container trucks.

• A parcel delivery service optimizes its packagedelivery on a street system.

• An electric utility locates where power outagesoriginate based on telephone calls received fromaffected customers.

• An environmental agency analyzes watersamples collected from streams to tracecontaminant flow.

• A regional transportation agency uses traffic datato plan future highway construction.

• A school district finds optimum bus routes topick up children and deliver them to school.

• A driver uses a mapping system with a GPSreceiver mounted in the car to find the best wayto get to a destination.

Chapter 8 • Linear modeling with networks • 129

NetworksYou can subdivide networks into two broad �

categories: transportation and utility.��

In a utility network, water and electricity � are channeled until delivery to the customer.�

�In a transportation network, cars and trains are �

autonomous objects that can move freely.� ��Transportation networks

Utility networks

Some transportation network tasks are:�• Calculating the shortest path between points�• Determining a trade area based on travel time�• Dispatching the closest ambulance�• Finding the best sequence to visit customers�• Routing a garbage truck efficiently

Some utility network tasks are:�• Establishing the direction of commodity flow�• Finding what is upstream of a point�• Closing switches or values to redirect flow�• Identifying isolated parts of the network�• Finding facilities that serve a set of customers��


The geodatabase has a dual representation of alinear system—the geometric network and thelogical network.

The geometric network is the set of features thatparticipate in a linear system. The geometricnetwork matches a view of a network as acollection of features.

A geometric network is associated with a logicalnetwork, which is a pure network graph consistingof edges and junction elements.

Together, these two representations of a networkprovide a rich model for storing and analyzinglinear systems.

THE GEOMETRIC NETWORK

A geometric network is a collection of features thatcomprise a connected system of edges andjunctions. An edge has two junctions and a junctioncan be connected to any number of edges.

Edge features can cross in two-dimensional spacewithout intersecting. An example is a bridge over aroad. This is called nonplanarity.

The features that represent edges and junctions arecalled network features. Only network features canparticipate in a geometric network.

A network feature class is a homogeneous collectionof one of these types of network features: simplejunction feature, complex junction feature, simpleedge feature, or complex edge feature.

More than one network feature class can representa given topological role in a geometric network. Anetwork feature class is associated with exactly onegeometric network.

Network features in a geometric network have allthe same characteristics as other features:

• You can create as many feature classes asnecessary for edges and junctions. You can addany attributes to these feature classes.

• You can define subtypes for major featureclassifications and apply default values, attributedomains, and split/merge policies on attributes.

• You can establish relationships among networkfeatures and any other feature or object.

• For advanced applications, you can extend anetwork feature class and create custom networkfeatures.

Network features have additional specializedbehaviors that preserve connectivity andautomatically update network elements.

THE LOGICAL NETWORK

Like a geometric network, a logical network is acollection of connected edges and junctions. Thekey difference is that a logical network does nothave coordinate values. Its main purpose is to storethe connectivity information of a network alongwith certain attributes.

Since edges and junctions in a logical networkcontain no geometry, they are not features, butelements. There is a one-to-one or one-to-manyrelationship between network features in ageometric network and network elements in alogical network.

A geometric network is always associated with alogical network. The logical network elements areautomatically updated when you edit networkfeatures.

The logical network does not directly appear in theArcInfo applications. Rather, you interact with thegeometric network. The logical network is the basisof the sophisticated behavior of the networkfeatures.

Reading the diagrams

There are conceptual diagrams throughout thischapter that show the relationship between thegeometric network and its logical network.

Many of the details about the logical network aresimplified—it is not necessary for the data modelerto know all of the internal implementation.

While the logical network is invisible in the ArcInfoapplications, understanding its basic concepts willgive you insight into building network models.

THE NETWORK MODEL


Junctions

Edges

Train stationsid geometry

Airportsid geometry

Riversgeometryid

Railroadsid geometry

idRoads

geometry

Two views of a networkYou can view a network as a collection of geographic objects such as rails, roads,

stations, and bridges and also as a pure network of edges and junctions.

ElementsFeatures

Network features can beorganized in any number of

network feature classes.

Network elements are stored in an edge tableand junction table with a connectivity tabledescribing how they connect.

You interact with the network through network features. When you add or remove a network feature in ageometric network, ArcInfo adds or removes the matching network elements. When you perform network

analysis, ArcInfo submits a solver to the logical network.

The geometric network and logical network are always synchronized.

Features from an arbitrary number of edge andjunction feature classes correspond to networkelements in the edge table and junction table.

A network featurecan be related to oneor many networkelements. This allowsa single feature torepresent a complexpart of a network.

A logical network hasno geometry or

coordinates. It is apure graph of how

junctions and edgesare connected.

The logical network is invisiblein ArcMap and ArcCatalog,

but it is the foundation for thegeometric network’s rich

model and high performance.

Airport

Airport

Factory Bridge

Navigable riverBridge

Railroad

Airline route

43

66

Train station

Navigable river

Railroad

Airline routeHighway

Logical networkGeometric network

A logical network is a pure graph of junction elementsand edge elements.

A geometric network is the representation ofgeographic features that comprise a network.

Geographic view Network view

The geometric networkmaintains relationshipsbetween connectedjunction features andedge features. When youmove a junction feature,the connected edgefeatures arerubberbanded.

You can define connectivityrules to define the validcombinations of connectedjunctions and edges in ageometric network.

The connectivity tablekeeps track of how edge

and junction elements areconnected.

Connectivityjunction

IDsadjacent junctions

and edges

Junctionfeatureclasses

Edgefeatureclasses

Logicalnetworktables

Highway


The centerpiece of a logical network is theconnectivity table, which describes how networkelements are connected.

A logical network contains the connectivity of thenetwork. The connectivity table lists all the adjacentjunctions to a given junction, along with the edgethat connects them.

j123

j126

j124

j125

id

j123

j124

j125

j126

geometry

Junction feature table

id

e1

e2

e3

geometry

Edge feature table

A geometric network contains the features that participatein a network. Feature classes contain either edgefeatures or junction features.

Geometric network

j123

j124

j125

j126

Junction

j124, e1

Adjacent junction and edge

j124, e1 j125, e2 j126, e3

j124, e2

j124, e3

Connectivity table

Logical network

e1

e2

e3

For every junction in the network, the connectivitytable lists the adjacent junctions and edges—junctions at the other end of the connected edge.

The connectivity table is how the geometric networkmaintains the integrity of the network.

The logical network also contains a junctionelement table and an edge element table.

A geometric network can have any number of participatingfeature classes. In this example, there is one junctionfeature class (Cities) and two edge feature classes thatconnect the junctions (Major rails and Truck routes).

Geometric network

id

t1

geometry

Truck routes

id

j1

j2

j3

j4

geometry

Cities

id

r1

r2

geometry

Major rails

j3

j1j2

j4

r1

r2

t1

r1

r2

10

11

12

2

2

Featureclass

FeatureID

ElementID

Edge element table

t13

j1

j2

j3

j4

0

1

2

3

1

1

1

1

FeatureClass

FeatureID

ElementID

Junction element table

The logical network tracks feature IDs by feature class.For each feature class and feature ID combination, thelogical network creates its own internal element ID.

Logical network

Connectivity table

0

1

2

3

1, 10

0, 10 2, 11 3, 12

1, 11

1, 12

Junction Adjacent junction and edge elements

The junction element and edge element tablesprovide a unique element ID that is a combinationof the feature class and the feature ID.

HOW FEATURES CONNECT


CONNECTIVITY RULES

In most networks, not all edges can connect to allother junctions. Also, not all edges can connect toall other edges through a specified junction. Forexample, a hydrant lateral in a water network canconnect to a hydrant, but not to a service lateral.Similarly, a 10-inch transmission main can onlyconnect to an 8-inch transmission main through areducer.

Network connectivity rules constrain the type ofnetwork features that may be connected to oneanother, and the number of features of anyparticular type that can be connected to features ofanother type.

Connectivity rules let you easily maintain theintegrity of the network features in a geometricnetwork. At any time, you can selectively validatefeatures in the database and generate reports as towhich features in the network are violating one ofthe connectivity or other rules. The following arethe connectivity rules for network features.

Edge–junction rule

This rule constrains which types of junctions canconnect to a type of edge.

No, a transmission main cannotbe connected with a meter.

Yes, a service tap can beterminated with a meter.

meter

service tap transmission main

meter

Meters can only connect to low-voltage lines.

Edge–edge rule

This rule establishes which combinations of edgetypes can connect through a given junction.

No, there is no reducer at thejunction connecting two pipes.

Yes, this reducer properlyconnects the two pipes.

10" pipe 8" pipe

reducer

10" pipe 8" pipe

no reducerat junction

Two pipes of different diameters can be connectedonly through a properly sized reducer.

Edge–junction cardinality

This rule lets you restrict the count (cardinality) ofedges that connect at a junction.

No, a switch cannot be theterminus of a line.

Yes, a switch can connect twothrough four lines.

switchswitch

A certain type of switch might be designed toaccept between two and four lines. You canprecisely define the acceptable range of lines thatcan be connected at a junction.

Default junction type

When you connect one type of edge to another,you can specify a default junction type to beinserted.

When you connect a 14.4-kV electric distribution line to a 28.8-kVline, a properly sized transformer can be automatically inserted.

step transformer

14.4-kV line 28.8-kV line

When a 14.4-kV line is added to an end-junction ofa 28.8-kV line, a step-down transformer with thecorrect electrical ratings is assigned to the junction.


Features can play four roles in a geometric network:simple edge, simple junction, complex edge, andcomplex junction.

Each feature class in a geometric network containsfeatures of one of these types. A geometric networkcan have multiple feature classes with the samerole.

This is a simplified portion of the geodatabase dataaccess model.

Junction-Feature

Network-Feature

EdgeFeature

Complex-Edge-

Feature

Simple-Edge-

Feature



There are two types of network features: junctionand edge. There are two types of junction features:simple and complex. There are two types of edgefeatures: simple and complex.

SUMMARY OF NETWORK FEATURES

A simple edge feature is associated with a singleedge in a logical network.

Network feature Network element

1 to 1

simple edge feature

A simple junction feature is a feature associated witha single node in a logical network.


1 to 1

simple junction feature

A complex edge feature is associated with anynumber of edges in a logical network. These edgesmust be arranged in a chain configuration.


1 to many

complex edge feature

A complex junction feature is associated with acollection of junctions and edges in a logicalnetwork. The edges and junctions are connectedand may be arranged in any topologicalconfiguration. These elements can be considered tobe an internal network represented by a complexjunction feature.


1 to manyPMH-7

complex junction feature

The complex junction feature must be implementedby writing a custom feature type. You cannot createa new feature class based on complex junctionfeatures without writing software code compliantwith the ArcInfo class extension framework.

Simple edge and junction features

These network features have a one-to-onecorrespondence with network elements. They aresuitable for simple parts of networks, but the edge-splitting scenario shows one limitation of simplenetwork features.

NETWORK FEATURES


Geometric network

j3

j1j2

j4

r1

r2

t1

Logical networkid

t1

geometry

Truck routes

id

j1

j2

j3

j4

geometry

Cities

id

r1

r2

geometry

Major rails

r1

r2

10

11

12

2

2

Featureclass

FeatureID

ElementID

Edge element table

t13

j1

j2

j3

j4

0

1

2

3

1

1

1

1

Featureclass

FeatureID

ElementID


id

e1

e2

e3

geometry

Water mains

diameter type

15

15

15

concrete

concrete

concrete

e1

e2

e3

10

11

12

13

14

1

1

1

Featureclass

FeatureID

ElementID

Edge element table

h1

h2

2

2

Simple edge and junction featureshave a one-to-one correspondence

between the feature in thegeometric network and the element

in the logical network.

Simple edge and junction features

Splitting a simple edge

id

h1

h2

geometry

Water mains

Service

In this example, there is one water main that servicestwo houses. In order for there to be flow from the watermain to the houses, two additional junctions need to beplaced on the main.

With simple edges, the only way to add junctions is tosplit the main into three separate edge features, becausea simple edge has a one-to-one correspondence withedge elements.

Water mains

Service

Service taps

id

e1

diameter

15

type

concrete

geometry

geometryid

t1

t2id

h1

h2

geometry

id

e1

e2

e3

diam

15

15

15

type

concrete

concrete

concrete

geometry

e1

h1

h2 h2

h1e1

e2

e3


Complex edge features

Imagine that your data has a municipal water mainthat runs for several hundred meters along a street.Along this main are service taps (junctions)connected to household service pipes. All your dataquery and maintenance functions ideally need totreat this main as one single feature.

But because your network analysis functions needto model flow from the main to the services, thelogical network needs to treat the portions betweeneach service tap as a single edge. Using simpleedges, this main would have to be broken intosections between each junction. So instead of

having one long main, there are now manyfragmented sections of the main, which complicatesdata query and maintenance.

With simple edges, a single feature has to be splitinto many features in order to connect otherfeatures with it. This may be undesirable for manydatabases, leading to fragmented databases andcomplex rules of behavior.

Complex edges solve the fragmentation problem byallowing junctions to be placed anywhere alongtheir length without creating new edge features. Ageometric network with complex edge featurescreates many edge elements for each feature.

For this complex edge, the geometric network creates three edgeelements from the one edge feature, assigning a sud-ID to each feature ( e1-1 , e1-2 , and e1-3 ).

Logical representation

id

e1

geometry

Water mainsdiameter type

15 concrete

id

t1

t2

geometryService taps

1

2

3

Sub-ID

e1

e1

e1

10

11

12

13

14

1

1

1

Featureclass

FeatureID

ElementID

Edge element table

h1

h2

2

2

1

1

Connectivity table

1

1

Sub-ID

t1

t2

0

1

3

3

Featureclass

FeatureID

ElementID


Water mains

ServiceService taps

Geometric network

t1t2

t1t2

id geometry

Logical network

Splitting a complex edge

e1geometrydiameter type

15 concreteid

id geometryh1h2

h2

t2

t1

h1

e1h1

e1-2

e1-3

e1-1

h2

e1

1

110

13

1410

12

0

1

Junction

–, 10

Adjacent junction and edge

0, 11 –, 12

1,11 –, 14

–,13


Complex junction feature

A complex junction feature corresponds to anynumber of edge and junction elements in thelogical network.

The best way to understand complex junctions is toimagine a switch cabinet in an electrical network.Switch cabinets are actually miniature networks,consisting of their own simple junctions and edges.

A complex junction is ideal for modeling networkswithin networks, as in the case of electricalswitches. A complex junction can contain anynumber of edges and junctions.

2021222324


Featureclass

FeatureID

ElementID

Edge element table

Sub-ID

1111

s1s1s1s1

1234

1011121314151617

2222

c1c2c3c4

1111

-----

s1s1s1s1s1

56789

Featureclass

FeatureID

ElementIDSub-ID

c1

c2

c3

c4

s1-5

s1-6

s1-7

s1-8s1-9

s1-1 s1-3

s1-2 s1-4

In the logical network,this switch is modeledwith four edge elementsand five junctionelements. You can writecode in a custom switchobject to manage theedges and junctionelements in the logicalnetwork.

Logical network

Complex junctions areused in electricalnetworks to representcomplex switches. Thisschematic of a complexswitch shows two simpleswitches (SW) and twofuses (F).

Schematic

SW1 F1

SW2 F2

In a geometric network, this switch is displayed as abox labeled with the type of switch, in this case “SW-2.”�Two wires enter and leave this switch. This switchcould be implemented as a complex junction feature.

Geometric network

Conductorfeature class

Switch feature classc1c2c3c4

id geometry

id type geometry

SW-2s1

c1

c2

c3

c4SW-2

Complex junction features


Pump station scenario

This example shows a pump station as a complexjunction. The pump station is displayed as a box.Inside this box is a miniature network containingthree valves, a check valve, a pump, a meter, and atee, all modeled as simple junctions within the onecomplex junction.

In ArcInfo, a developer creates complex junctionsas a custom feature, implementing the rules on howthe junction is stored in the logical network.

P STransmission main

Hydrant

Isolation valveHydrant lateralTap

Service lateral

Meter vault

Pump station

Tee

Pump

Valve 2

Valve 3

Valve 1

Meter

Tee

Checkvalve

Geometric network

Logical representation

This is an example of a feature datasetcontaining a geometric network for waterdistribution.

The pump station is acomplex junction object,containing many simplejunctions and simple edges.


Networks fall neatly into one of two operationalcontexts, utility network or transportation network.In a transportation network, the commodities thatflow through the network—automobiles—have a“will of their own.” The driver of the automobiledecides how they will flow through the network. Ina utility network, the commodity that flows throughthe network—water, electricity, oil—has no will ofits own. The network imposes flow direction by itsconfiguration of sources, sinks, and switches.

Transportationnetwork

Utilitynetwork

In utility network applications, the direction ofcommodity flow along edges needs to be anintrinsic part of the network. For example, it isusually a bad thing to have water flowing in fromboth ends of a solitary pipe—the water is eithergoing to back up or the pipe is going to burst.

If the geometric network is used for operationaldecision making, such as whether to close a switchor open a valve, you have to know if the decisionwill result in incorrect flow. In analysis, it is usuallya requirement to know what features aredownstream (with the flow) or upstream (against theflow) of some location.

A geometric network has a method to establish flowdirection. This method decides how commoditiesflow in the network based on the currentconfiguration of sources and sinks and the enabledstate of each feature. The result of this method isto align the direction that commodities flow along

each edge, either with the direction of the feature oragainst the direction of the feature, relative to itsdigitized direction.

All line features have an implicit direction ofdigitization, which is the x,y coordinate order. This isan example of a simple stream network. All waterflows to the discharge point. Water flows oppositethe digitized direction of feature e1, but with thedigitized direction of features e2 and e3.

Logical network

The establish flow direction method on a geometricnetwork populates the flow direction property ofeach edge element in the logical network. There aretwo possible values: flow is against digitizeddirection or with digitized direction. This flowdirection information is critical for many applicationsand is used when flow direction is to be displayedon a geometric network.

Edge element table

Geometric network

Discharge point

Edge feature class

Digitized direction of feature

Direction of flow along feature

id

e1

e2

e3

geometry

e1

e2e3

Featureclass

FeatureID

ElementIDSub-ID

1

1

1

e1

e2

e3

1

1

1

10

11

12

Flowdirection

Against

With

With

SOURCES AND SINKS

In a utility network, sources and sinks are used indetermining flow direction. Any junction featureclass can take on the ancillary role of a source or asink. A source is a junction from which acommodity flows, such as a well-head pump. Asink is a junction where all commodity flowterminates, such as a wastewater treatment plant.

NETWORK FLOW


When you build a geometric network, you saywhether or not features in a junction feature classcan assume this ancillary role. If they can, theeditor can be used to specify whether an individualjunction within the feature class is either a sourceor sink.

Geometric network

e1

Sink

j1

e2

j2

Source

id

e1

e2

e3

geometry

j1

j2

Source

Sink

Ancillary roleid geometry

Flow junctions feature class

Junction features can have an ancillary role of source,sink, or neither. The role is stored in an attribute of thefeature class, which is accessed by the establish flowdirection method.

Edge feature classe3

The ancillary role field determines if a junctionfeature and element are a source or sink.

DISABLED FEATURES

All features participating in a network have anenabled/disabled state. Features that are disabled donot participate in network flow: nothing flows intoor out of the feature. Disabled features are usefulfor representing open electrical switches or closedvalves.

Sources, sinks, and the enabled/disabled state allaffect how flow is established in a network.

All line features have an implicit direction ofdigitization, which is the x,y coordinate order. This isan example of a simple stream network. All waterflows to the discharge point. Water flows oppositethe digitized direction of feature e1, but with thedigitized direction of features e2 and e3.

Logical network

The establish flow direction method on a geometricnetwork populates the flow direction property ofeach edge element in the logical network. There aretwo possible values: flow is against digitizeddirection or with digitized direction. This flowdirection information is critical for many applicationsand is used when flow direction is to be displayedon a geometric network.

Edge element table

Geometric network

Discharge point

Edge feature class

Digitized direction of feature

Direction of flow along feature

id

e1

e2

e3

geometry

e1

e2e3

Featureclass

FeatureID

ElementIDSub-ID

1

1

1

e1

e2

e3

1

1

1

10

11

12

Flowdirection

Against

With

With

INDETERMINATE FLOW

It may not be possible to establish flow directionfor an edge. This only occurs when the sources,sinks, and disabled features do not give enoughinformation. When flow direction cannot beestablished for an edge, it has indeterminate flow.

Indeterminate flow occurs when the establish flowdirection method cannot determine which directioncommodities flow in a network.


Geometric network

It may not be possible to determine the direction offlow given a configuration of sources, sinks, andenabled features. In this example, it is impossible todetermine the flow through edges e1 and e2 becausethey form a cycle. Changing junction j2 to a sourcewould break the cycle.

e1

Sink

e2

j1

j2

Logical network

The establish flow direction method will write“Indeterminate” as a flow direction when the flowdirection cannot be established.

Featureclass

FeatureID

ElementID

Edge element table

Sub-ID

1

1

1

e1

e2

e3

1

1

1

10

11

12

Flow direction

Indeterminate

Indeterminate

With

??

?

?e3

UNINITIALIZED FLOW

When a flow is isolated because the edges aredisconnected from the rest of the network (that hasflow), flow is said to be uninitialized.

Unreached edge features

When establishing flow direction, edge features may beunreached because they are disconnected from therest of the network. In this example, the unreachededges are disconnected because one of the junctionfeatures–a valve–is disabled.

Source

Disabled junction feature

WEIGHTS

Edges and junctions can have any number ofweights associated with them. Weights are typicallyused to store the cost of traversing across an edgeor through a junction. A typical weight is the lengthof the edge. Weights are created from field valueson the edge and junction feature classes.

Pipes

Edge and junction features canhave any number of weightsassociated with them.

Geometric network

Logical network

Weights are stored with the logical network.

e1

e2

e3

e4

id

e1

e2

e3

e4

geometrydiameter length

15

15

15

8

55.1

61.0

28.7

24.9

e1

e2

e3

e4

0

1

2

3

1

1

1

1

Featureclass

FeatureID

ElementID

Edge element table

1

1

1

1

Sub-ID Diameter Length

15

15

15

8

55.1

61.0

28.7

24.9

Weights are stored with the logical network so thatanalysis programs can access them efficiently.When a weight value is modified on a feature table,it is automatically updated in the logical network.

Any numeric field can be a weight. Determiningwhich fields should be weights depends entirely onthe types of analyses you wish to perform.


In ArcInfo, network analysis is a procedure thatnavigates through the connectivity of the network toyield some meaningful result, such as finding allelements upstream of a point or the shortest pathbetween two points.

There are other ways to analyze networks, ofcourse. For example, you could use the basicselection tools found in ArcMap to select edgefeatures and then calculate statistics about them,such as the total edge length by type of edge. Thisis certainly a valid analysis on a network, but it isnot a “network analysis” because the networkconnectivity is not involved.

Solvers

A program that performs network analysis is calleda solver, because it solves a problem, such asisolating flow to an edge by turning off a set ofvalves. Inputs to this example flow isolation solverwould be the logical network, the edge to isolate,and the set of junctions that are valves. The outputwould be the set of valves to turn off. There are norules about the inputs and outputs of solvers,except that input always includes a logical network.

Solvers have user interfaces for specifying inputsand reporting outputs. Collections of solvers thatperform similar tasks can usually be plugged into acommon user interface framework. For example,the ArcInfo trace solvers are all accessed through acommon toolbar. ArcMap is naturally part of theuser interface for a solver, because ArcMap iswhere you graphically identify solver input, such asstart points for a trace.

There are almost an infinite variety of solvers forthe many types of network analyses. The ArcInfostrategy is to provide a rich suite of solvers thataddress the more common types of problems. Forless common types of network analyses, developerscan create solvers using any programming languagethat can access the ArcInfo components.

NetFlags

A NetFlag is a location on a network. Solvers useNetFlags to represent a multitude of real-worldobjects, such as stops for a shortest path, startpoints for tracing, locations of valves, locations of

ANALYSIS ON A NETWORK

services, and so on. NetFlags are not part of alogical network. They are used to describe anylocation in a network.

There are two kinds of NetFlags: EdgeFlags andJunctionFlags. NetFlag properties include the LogicalNetwork element’s feature class, feature ID, andfeature sub-ID. An EdgeFlag additionally includesthe percent along the edge element. This means thatan EdgeFlag can fall anywhere along the edge, fromzero percent (the from-junction) to 100 percent (theto-junction).

80%

25%

EdgeFlag

JunctionFlag

25%

An EdgeFlag has a percent alongan element.

digitizeddirection

NetFlags are used to describe any location on anetwork. Examples include places to visit on ashortest path, the origin of a trace, a warehouse ora service center, or a valve, switch, or transformer.Solvers rely heavily on NetFlags to describe inputparameters.

Barriers

Barriers are used by solvers to represent disabledlogical network elements. Barriers do the same jobas setting an element’s enabled/disabled state todisabled, except that barriers are not stored with thelogical network—they are known only to the solver.Barriers are just a way to temporarily disableelements. Barriers are either edge or junctionelements.

There are four methods to capture and representbarriers to a solver. A well-designed solver willallow all four methods.


These are ways to set barriers to a solver:

• You can interactively add simple barriers.

• You can use the features in your selection set.

• You can disable feature classes.

• You can apply a weight as a filter.

TRACING

The ArcInfo network solvers currently work with aclass of utility network problems termed tracing.Future versions of ArcInfo will include moresolvers.

Tracing means to follow the flow in a network untilsome condition is met. When you hear problemsexpressed as “search against the flow until you finda transformer,” or “follow the flow upstream to thefirst discharge point,” or “trace upstream and findall valves,” you are most certainly looking at using atrace solver to find the answer.

The ArcInfo trace solvers include upstream trace,downstream trace, isolation trace, and path trace.

WEIGHTS

Choosing which edge or junction attributes shouldbecome weights in the logical network depends onyour collection of solvers. It is of no use to add aweight to a network if there are no solvers that canuse it. For example, trace solvers typically do notuse any weights—only the connectivity informationfound in the logical network.

For example, suppose you have a water distributionnetwork with a numeric attribute containing thepipe manufacture ID. There is no need to add thisattribute unless you have a solver that can use it.Even if you had a shortest path solver, it would notmake sense to find the shortest path based onmanufacturer ID.

But suppose you had a solver that could return alljunctions that share edges of certain characteristics.In this case, you may want to use this solver to findall junctions where pipes from manufacture 100connect with manufacture 151. In this case, it mightmake sense to add manufacturer ID as a weight.

Below is a table of just a few possible weightattributes and the types of solvers that would usethese weights.

Length of edge

Diameter of pipe

Impedance (electrical resistance)

Time to traverse an edge

Number of lanes on a street

Road classification

Miles per hour

Hazardous material route

Toll (cost to use a road)

Shortest path solvers. Many solvers have a need for length.

Solvers that calculate pressure or head in a network.

Calculating voltage drop in an electrical network.

Shortest path solvers.

Calculating traffic capacity or congestion on a street.

Used to describe network hierarchy in hierarchical shortest path solvers.

Used with a shortest path solver that allows dynamic calculation of weights.

Useful as a filter—find a path only on hazardous material routes.

Shortest path solvers based on actual cost.

Used forWeight description


Once the state of the networkis set, you specify the type ofsolver you want to execute.

Solver output has several formats, such as asingle number. Most typically, solvers outputcollections of elements and NetFlags.

A solver can display its trace result toa map through a solver renderer,

which determines the set of networkfeatures that match the network

elements in the trace. You can setseveral cartographic parameters to

control the appearance of the trace.

draws featurescorresponding tonetwork elementsfound from solver

solverrenderer

solverTrace Upstream

Trace UpstreamTrace DownstreamTrace ConnectedTrace Common AncestorsTrace LoopsTrace Path

logical network

NetFlag collections

simple barriers

disabled feature classes

selection sets

weight filters

geometric network

solver parameters

solver input

solver parameters

edge and junctionelement collections

NetFlag collections

solver output

In ArcMap, you caninteractively set the state of anetwork through theplacement of barriers, filters,and NetFlags by disablingcertain feature classes in thegeometric network, bycreating a selection set, andby setting solver parameters.

Network solvers


Network related objectsin the geodatabase

Feature-Dataset

Feature

Geometric-Network

Junction-Connectivity-

Rule

Edge-Connectivity-

Rule

Junction-Feature

Connectivity-Rule

Rule

1..*

1..*

Network-Feature

EdgeFeature

Complex-Edge-

Feature

Simple-Edge-

Feature

ObjectClass

Feature-Class

Graph



0..1

A junction that is represented byone and only one junction

element in the logical network.

A junction that is represented bymany junction and/or edge elementsin the logical network.

The endpoints of edges wherecommodities flows from oneedge to another edge.

An edge that is represented by oneand only one edge element in the

corresponding logical network.

An edge that is represented by manyedge elements in the logicalnetwork.

Lines through whichcommodities flow, such aspipelines, transmission lines,and roads.

A rule that specifies what edgescan connect to each other.

A rule that specifies whatedges can connect to ajunction.

A connectivity rulespecifies which networkfeatures can beconnected with anotherbased on subtype andattribute value.

A graph represents a setof topologically relatedfeature classes.

A network of edges andjunctions that have geometryand can be mapped.


creates

is associated with

is composed of

1..*

CreateableClass

AbstractClass

InstantiableClass

An abstract class is a specification of the methods and propertiesto be inherited by other classes. You cannot create objects froman abstract class.

You can directly create objects from a createable class with astatement like “Dim as New <object>”.

You can indirectly create objects from an instantiable class bycalling a method of another class.

Sample multiplicities: “1” or “ ” is one, “0..1” is zero or one, “ * ” iszero to any integer, “1..* ” is one to any integer.multiplicity

This is a simplified UML diagramshowing selected parts of thegeodatabase data access objects thatrelate to geometric networks.

Only network features canparticipate in a geometricnetwork.


Network object model

Network

NetWeight

ForwardStar

Network-Loader

EdgeFlag

NetFlag

JunctionFlag

Street-Network

Utility-Network

Feature-Workspace

Converts feature class and featureIDs to and from element IDs.

Returns all the adjacent elementsgiven a junction and edge element.Used extensively when writingsolvers.

Provides information about aweight.

Creates a geometric and logicalnetwork from line and point feature

classes.

A network where flow is directed.You can establish a flow direction

on this type of network.

A network where flow is undirected.

This is the logical network. Thisobject provides for examination and

editing a logical network.

A NetFlag is a location on anetwork. NetFlags are used bya NetSolver.

A NetFlag that occurs at a junction. A NetFlag that occurs somewherealong an edge. The position ismeasured as the percent along theedge.

This is a simplified UML diagram extracted from theArcInfo object model showing the network objects.

NetElement-Description

Specifies weights to be used ina solver.

NetSolver-Weights

Implements trace solvers.

Flow-Direction-

Solver

Creates barriers of individualelements.

NetElement-Barriers

Creates barriers from selectedsets of features.

Selection-Set-

Barriers

147

9 Cell-basedmodelingwith rasters

Some geographic phenomena can berepresented best as locations on a grid withvalues. This data structure is called a raster.In this chapter:

• Representing geography with rasters

• Using raster data

• Raster data model

• Raster display and analysis

• The spatial context of rasters

• Raster formats

• Raster object modelMosaicked image of Mars on a simple cylindrical projection. MarsGlobal Surveyor, Malin Space Science Systems and JPL/NASA, 1999.


REPRESENTING GEOGRAPHY WITH RASTERS

A raster is a rectangular array of equally spacedcells, which taken as a whole represent thematic,spectral, or picture data. Raster data can representeverything from qualities of a land surface such aselevation or vegetation, to satellite images, scannedmaps, and photographs.

The raster data format is very simple but supports arich variety of data types. This chapter discusses theways that geographic data is represented in rasters.

FORMS OF RASTER DATA

Raster data can be entered into a GIS throughimaging systems or calculated from other data.

Satellite imagery

Acquiring satellite images is a cost-effective way tomap a sizeable part of the world at small tomoderate map scales.

Satellite images are perhaps the best system in a GISto capture temporal changes in the landscape. Youcan compare and analyze scenes of the same areafrom different seasons or years.

Images can show a scene in color or black andwhite. Color information is stored as an RGBcomposite pixel value or as a set of raster bandsrepresenting several or many colors.

Some satellite imagery services allow you to orderfresh images, only days or hours old. These are animportant asset for managing environmental eventssuch as floods or forest fires.

Aerial imagery

To create detailed maps, airplanes with special large-format photographic and digital cameras record stripsof images that overlap to cover an area.

These images are then rectified for scale distortioncaused by the surface’s shape.

Scanned maps

Sometimes, the best basemap is a published mapthat is scanned. These can be assigned ageographic reference, so that the scanned map canbe precisely registered with other geographic data.

The USGS (United States Geological Survey)quadrangle maps are a good example of maps thatare scanned. These maps show terrain, placenames, rivers, roads, and major features.

Pictures

Besides images of land, rasters are also used forphotographs of features. Photographs of featurescan be an important augmentation to theinformation presented in a map.

Chapter 9 • Cell-based modeling with rasters • 149

Converted data

Rasters can also be generated from other datasources, such as feature datasets and TIN datasets.Analysis results done from raster surface studiescan be used to make a slope map. An image classi-fication can be used to make a land-cover map.

TYPES OF RASTER DATA

There are two general categories of raster data:thematic data and image data. Thematic data can beused for geographic analysis of land use. Imagedata is used for basemaps to other geographic dataand to derive thematic data.

Thematic data

The value of each cell (or pixel) in a raster can be ameasured quantity or classification. When drawn,these rasters are thematic maps.

Spatially continuous data

The values of raster cells can represent a measuredquantity such as elevation, pollution concentration,or rainfall. The value from one cell to anothervaries slightly, and collectively, these values canmodel some type of surface.

Cell values for spatially continuous data represent asampled quantity at cell centers.

Spatially discrete data

Values for raster cells can represent a category orclassification of data, such as land ownership typeor vegetation type. The value from one cell toanother is typically either identical or changesabruptly. This type of data appears as a set of zonalregions with common values, such as land-usemaps or forest stands.

Cell values for spatially discrete data represent aclassification that applies to the full area of a cell.

Image data

The preponderance of raster data is captured byimaging systems mounted on satellites andairplanes.

Spectral and picture data

Imaging systems record rasters based on lightreflectance values at one or many bands of theelectromagnetic spectrum.

Picture data usually captures the red, green, andblue portions of the spectrum for display on amonitor or map, but certain satellite images capturemany bands that are used to analyze surfacegeology and vegetation.


USING RASTER DATA

Rasters are used for display and analysis. Thesetopics illustrate just a few of the uses of rasters.

Basemap

Often, rasters are used as a visual backdrop on amap. They are the bottom layer on which vectorand surface data is drawn.

For example, you can draw road features on top ofa raster image of a city and instantly see what partsof the road network need updating.

Using a good raster image as a basemap layer addsdepth to the map and enhances the user’s trust inthe map. A raster basemap is a good check on mostother geographic data.

Land-use scenario

Raster data is ideal for modeling and mapping landuse and land-use change. Most land-use studiesbegin with satellite or aerial imagery that isinterpreted and then categorized into classes suchas urban, agriculture, and deciduous forest.

Over time, these studies are repeated, anddifferences between years can be analyzed.

Hydrological analysis

Terrain information is commonly available in araster form with elevations for cell values. This iscalled a digital elevation model (DEM).

Raster GIS tools let you to determine the directionof water flow across the landscape, downstreamaccumulation of precipitation, and the delineationof drainage basins or watersheds.

This model is the basis for doing hydrologicanalysis such as runoff prediction from a storm andwhich structures are at risk of flooding. Thisinformation is useful for mapping floodplains andfor determining flood insurance rates.

Environmental analysis

Because data such as land cover, vegetation type,and terrain are commonly stored as rasters, mostenvironmental analysis involves raster data.

Raster GIS analysis tools have evolved to use suchdata to solve problems at many scales. These rangefrom continental forest succession as a result ofglobal warming to local wildlife habitat changesresulting from urbanization.


Terrain analysis

Digital elevation models contain elevations for cellvalues. ArcInfo contains raster analysis tools tostudy the visibility, slope, aspect, and curvaturecalculations that are often used as part of a largestudy, such as land-use planning or site selection.

You can pose questions such as, “Find all thelocations between 2,500 and 5,000 feet in elevation,facing south or southeast, on a slope of less than12 percent, with 3 miles of visibility.”

ArcInfo also includes display functions for digitalelevation models, an example of which is analytichillshading, which produces a realistic view ofterrain. These are some maps that show surfacedisplay methods.

For each cell in a digital elevation model, ahillshade map draws shades that simulate theillumination of a surface, based on the anglebetween the sun and the slope of the local surface.

This map shows the slope of a terrain. The red cellsshow steep areas and the green cells show flat areas.

This map shows elevation by shaded colors. Thegreen cells show lower elevation. The red, pink,and white cells show higher elevation.

This map shows elevation combined withhillshading. The combination of these displaymethods creates an attractive map thatsimultaneously shows heights and the surfaceshape.


RASTER DATA MODEL

Rasters are made of cells. A cell is a uniform unitthat represents a defined area of the earth, such asa square meter or square mile.

Each raster cell has a value that represents aspectral reflectance or a characteristic at thatposition, such as a soil type, census tract, orvegetation class. Additional values of the cell canbe stored in an attribute table.

The size chosen for a grid cell of a study areadepends upon the data resolution required for themost detailed analysis. The cell must be smallenough to capture the required detail, but largeenough so that computer storage and analysis canbe performed efficiently. The more homogeneousan area is for critical variables such as topographyand land use, the larger the cell size that can beused with accuracy.

CELL ATTRIBUTES

The value associated with a cell defines the class,group, category, or measure at the cell position. Cellvalues are numbers: integer or floating point.

Cell locations with the same value belong to thesame zone. Cells of the same zone do not have tobe connected.

When an integer value is used for cells, it may be acode for a much more complex identification. Forexample, the value four may equate to single-familyresidential parcels on a land-use grid. Associatedwith the value of four might be a series ofattributes, such as the average commercial value,average number of inhabitants, or census code.

There is usually a one-to-many relationshipbetween the grid cell values (or codes) and thenumber of cells that are assigned the code. That is,there might be 400 cells with the value four(representing single-family residential) and 150 cellsassociated with the value five (representingcommercial zoning) on the land-use grid.

The code value occurs many times in the raster, butonly once in the attribute table, which storesadditional attributes for the code. This designreduces storage and simplifies updating. A singlechange to an attribute can be applied to severalhundred instances of that value.

TYPES OF DATA

Each cell in a raster has one value. The cell valuesin a raster can represent one of the following fourgeneral types of data.

Nominal data

A value of nominal data identifies one entity fromanother. These values establish the group, class,member, or category with which the geographicentity at the position of the cell is associated. Thesevalues are qualities, not quantities, with no relationto a fixed point or a linear scale. Coding schemesfor land use, soil types, or any other attributequalify as a nominal measurement.

Ordinal data

A value of ordinal data determines the rank of anentity versus other entities. These measurementsshow place, such as first, second, or third, but theydo not establish magnitude or relative proportions.You cannot infer a quantitative difference, such ashow much an entity is larger, higher, or denser thanthe others.

Interval data

A value of interval data represents a measurementon a scale such as time of day, temperature inFahrenheit degrees, and pH value. These values areon a calibrated scale but are not relative to a truezero point. You can make relative comparisonsbetween interval data, but their measure is notmeaningful when compared to the zero point of thescale.

Ratio data

A value of ratio data represents a measure on ascale with a fixed and meaningful zero point.Mathematical operations can be used on thesevalues with predictable and meaningful results.Examples of ratio measurements are age, distance,weight, and volume.


Inside a raster

23

29

31

37

41

43

7

18

10

18

4

7

Fir

Juniper

Aspen

Piñon

Cottonwood

Walnut

400

410

420

500

510

600

The attribute table

Rasters are two-dimensional arrays of cells (or pixels). Theheight and width of each cell are fixed and the same. A raster

spans a rectangular area.

Each cell has a value. This value can represent many qualitiesof a location, including reflectance, color, precipitation, and

elevation.

Rasters have an integer coordinate space. You can determinethe coordinate of a cell by counting columns from the left and

rows from the top. Row and column values begin with 0.

cell (pixel)

width

height

Types of data represented in cells

Nominal data

Ordinal data

Interval data

Ratio data

FirJuniperAspenPiñonCottonwoodWalnut

very goodgoodmoderatepoor

700–709710–719720–729730–739740–749750–759

0.0–10.010.1–20.020.1–30.030.1–40.040.1–50.0

Rasters that have integer valuedcells can be defined with an optionalattribute table, which recordsattributes for each unique cell value.

You can add custom fields to theattribute table.

21.1 17.3 17.2 18.1

18.5 16.2 17.3 19.1

21.0 19.1 19.4 19.2

26.3 23.1 21.6 20.5

Interval and ratio datapresent continuousphenomena and areusually measured withreal cell values.

24 23 18 16 20 19

18 14 16 17 19 20

21 17 17 18 22 18

18 16 17 19 24 19

21 19 19 19 22 22

26 23 21 20 18 21

Nominal and ordinaldata representdiscrete categories.They are bestrepresented withinteger cell values.

The data stored in a raster can be categorized as one of these types.

Nominal data values are categorized and have names.The data value is an arbitrary type code. Examples aresoil types and land use.

Ordinal data values are categorized, have names, andthe value is in a numerical rank. Examples are landsuitability classifications and soil drainage rank.

Interval data values are numerically ordered and theinterval difference is meaningful. Examples are voltagepotential and difference in concentration.

Ratio data values measure a continuous phenomenonwith a natural zero point. Examples are rainfall andpopulation.

Value Count Type Code

x,y coordinates are (5,3)

row

0 1 2 3 4 5 6

0

1

2

3

4

raster

column

Cell values can beintegers or real numbers.

19.4

19

nodata

Cells can also have a NODATA value torepresent the absence of data.


s

RASTER DISPLAY AND ANALYSIS

The rendition of rasters

Multiband rasters are often displayed as red-green-blue composites. This band configuration is common because these bands can be directly displayed on computer displays, which employ a red-green-blue color rendition model.

Raster datasets have one or manybands. In multiband rasters, a band

represents a segment of the electromagnetic spectrum that has

been collected by a sensor.

Redband

Greenband

Blueband

Red-green-bluecomposite

Attribute valuesrange from 0 to 255

in each band

255

0

Displaying multiband rasters

Electromagnetic spectrum

band 1

band 2

band 3

Bands often represent a portion of the electromagneticspectrum, including ranges not visible to the eye—theinfrared or ultraviolet sections of the spectrum.

A raster can have one or many bands. The cell values of rasters can be drawn in a variety ofways. These are some of the ways to display rasters by cell values.

Monochromeimage

0 0 0 0 1 1

1 0 0 1 1 0

1 0 1 1 0 0

0 0 0 1 1 0

1 1 0 0 0 1

0 1 1 1 0 0

0

251

41

86

118

141

187 236

201

16 25532

66

126

124 183

191 198

0

243

68

76 124

162

170

212

251 10

255

56

68

124

132

152

218

234

00 1 255

Grayscaleimage

Display colormapimage

1 3

42

5

1

3

4

2

5

1 3

4

2

5

1

3

42

5

1

3

4

2

5

1 3

42

5

1 34 2

5

2

1

3

4

2

5

Colormap

red green blue

64

255 0

128

3232

0

255

255 128

25500

128

255

In a monochrome image, each cellhas a value of 0 or 1. They are oftenused for scanning maps with simplelinework, such as parcel maps.

In a grayscale image, each cell has avalue from 0 to 255. They are often used for black-and-white aerialphotographs.

One way to represent colors on animage is with a colormap. A set ofvalues is arbritrarily coded to match adefined set of red-green-blue values.

Cell values in single-band rasters can be drawn in these three basic ways.

Displaying single-band rasters


raster operators

Calculations with rasters

input raster

raster operator

input raster

resultant raster

yields

Raster operators

When you are studying an area, you may wantto apply a suitability analysis. To do this, youwould select rasters with values such asrainfall, soil alkalinity, and insolation and applya series of operators according to your formulafor suitability. Operators can be arithmetic,Boolean, relational, bitwise, combinatorial,logical, accumulative, and assignment.

Map calculation

Map query

Mathematical operations can be applied to two rastersand the result is in the output raster. Functions include+, –, /, *, Log, Exp, Sin, Cos, and Sqrt.

You can apply Boolean and logical operators on two rastersto create an output raster with true/false values. Operatorsinclude And, Or, XOr, Not, >, >=, =, <>, <, and <=.

4 2

1 3

3 4

1 1+ =

7 6

2 4

sand clay

sandclay

dry dry

wet wet+ =

true false

false false

Raster functions

Local functions perform a calculation on a single cell at a time.The neighboring cells do not influence the result. The functionscan be applied to one raster or several overlaid rasters. Localfunctions include trigonometric, exponential, logarithmic,reclassification, selection, and statistical functions.

Focal functions perform a calculation on a single cell and itsneighboring cells. A neighborhood can be a rectangle, circle,annulus (doughnut), or wedge. These functions can return themean, standard deviation, sum, or range of values within theimmediate or extended neighborhoods.

Zonal functions perform a calculation on a zone, which is a setof cells with a common value. The cells that form a zone can bediscontinuous. There are two categories of zonal functions:statistical and geometric. These functions include area, centroid,perimeter, ranges, and sum calculations.

Global functions perform a computation on the raster as awhole. Examples are the calculation of Euclidean distances,weighted cost distances, and watershed delineation.

Localfunction

Focalfunction

Zonalfunction

Globalfunction

There are many raster functions. Each can accept one or many rastersas input and generate one or several rasters with the calculated results.

Input raster(s)

applied to araster functionyields

resultant raster(s).


THE SPATIAL CONTEXT OF RASTERS

When a raster is captured with devices like satelliteimaging systems or desktop scanners, the raw datais just rows and columns of cells. In order to usesuch data in a GIS either to draw on the screenwith other data or overlay in an analysis operation,the data must be in a common coordinate system.This is a real-world coordinate system.

The rows and columns of a raster are alwaysparallel to the x- and y-axes of the coordinatesystem.

Georeferencing

Georeferencing is the process of establishing arelationship between the raster’s (row, column)coordinate system, sometimes called image space,and a real-world (x,y) coordinate system, calledmap space. A similar transformation process occurswhen establishing the relationship of feature data inone coordinate system (i.e., digitizer units) toanother map coordinate system.

You can define a transformation to register the rasterto real-world coordinates. This lets your raster in thesame space as your other geographic data, such asvector features or a surface in a TIN dataset.

To georeference a raster, it needs to be registered toa coverage, map, or set of coordinates that are inmap space. The registration process is normally aninteractive one where common locations areselected both in the raster and the other geographicdataset. For imagery, these are normally things likeroad intersections that are easily identifiable in bothdatasets. Once the common locations are selected,a polynomial transformation is built to model thescale, rotation, and skew between the twocoordinate systems.

The georeferencing information is stored internallyto some raster formats, such as the ESRIARC GRID™ or ERDAS IMAGINE®, or in externalfiles such as the raster auxilliary file (.aux) or theworld file for other formats such as JPEG or BMP.

Using this information, the raster can betransformed on-the-fly and drawn in the map spaceof your other data. If you have also stored the mapprojection information, it can also be projected intoother coordinate systems.

Rectification

To align an image axis with a map-space axis, animage must be rectified by resampling it based onthe transformation built during the registrationprocess. In resampling, a mesh is overlaid on theraster and a value is assigned to each cell accordingto the center’s proximity to the values of the centersof the cells in the rotated raster. The values assignedto the output raster will be determined by the typeof resampling: nearest neighbor assignment, bilinearinterpolation, or cubic convolution.

You may wish to rectify a raster to remove skew orrotation, to orient its cells orthogonally to the maporientation. The primary reason not to rectify isbecause any such resampling of a raster will inducea small amount of error. This amount of error is notsomething you can see but can be important inmultispectral analysis where minute differences incell values can be significant.

Raster pyramids

Large rasters can display quickly if pyramids havebeen created for them. They often contain moreinformation than can be displayed on the screen. Ifpyramids are not present, then the entire rasterdataset must be investigated, and many calculationsmust be made to choose which subset of data cellsis sent to the display. Pyramids are a way to storereduced-resolution copies of the raster, and bychoosing a resolution that is similar to the amountof display area, there are fewer cells to investigateand fewer calculations, which therefore decreasesthe display time.


Referencing cell positions

Image to world affine transformation

x' = Ax + By + Cy' = Dx + Ey + F , where

x is column count in image space.y is row count in image space.x' is horizontal value in coordinate space.y' is vertical value in coordinate space.

A is width of cell in map units.B is a rotation term.C is the x' value of the center of upper-right cell.D is a rotation term.E is negative of height of cell in map units.F is the y' value of the center of upper-right cell.

Rasters are stored as arrays of cells (pixels)and can be displayed on the map’s

coordinate system. Rasters of geographicareas have a display transformation that

converts cell units to map coordinates.

Six parametersdefine how araster’s rows andcolumns transformonto mapcoordinates. Coordinate space

x'

y' A

E

map units: meters, feet, other

CImage space

display units: pixel

columns

rows

x

y

F

When you add raster datasets in ArcInfo, you have the option tocreate associated pyramids. They comprise a set of rasters that

are progressively downsampled by a factor of two.

As you zoom out and the raster cells grow smaller than thescreen display pixels, ArcMap will select one of the pyramided

rasters to draw.

The purpose of pyramids is to optimize display performance.

313 318 325 323

317 323 328 326

315 319 321 323

20 25 30 20

35 40 35 25

50 45 40 35

Value applies to center point of cell

Value applies to whole area of cell

For certain types of data, the cell valuerepresents a measured value at the center pointof the cell. An example is a raster of elevationvalues.

For most data, the cell value represents asampling of a phenomenon and the value ispresumed to represent the whole cell square.


You can display most industry-standard rasterformats in ArcInfo. This is a summary of thoseformats with a description of their attributes.

• ADRG (ARC Digitized Raster Graphics) dataconsists of raster images scanned and distributedon CD–ROM by the United States NationalImagery and Mapping Agency (US NIMA). ADRGis geographically referenced using the equal arc-second raster chart/map (ARC) system in whichthe globe is divided into 18 latitudinal zones.

• CADRG (Compressed ARC Digitized RasterGraphics) data is the same as ADRG data, but iscompressed with a nominal ratio of 55:1.

• CIB (Controlled Image Base) images arepanchromatic (grayscale) images that have beengeoreferenced and corrected for distortion dueto topographic relief. They are distributed by USNIMA and are commonly used as basemaps.

• DTED Level 1 & 2 is elevation data usuallyproduced by US NIMA.

• ERDAS® 7.5 GIS images are single-band thematicimages produced by the ERDAS 7.5 (or earlier)image-processing software.

• ERDAS 7.5 LAN images are single- or multibandcontinuous images produced by the ERDAS 7.5(or earlier) image-processing software.

• ERDAS raw data is used to read and displayuncompressed, band interleaved by line, bandinterleaved by pixel, and band sequential imagedata. Through an ASCII file that describes thelayout of the image data, black-and-white,grayscale, pseudocolor, and multiband imagedata can be displayed without translation into aproprietary format.

• ERDAS IMAGINE files are produced usingIMAGINE image-processing software. IMAGINEfiles can store both continuous and discretesingle-band and multiband data.

• ER Mapper files are produced using the ERMapper image-processing software.

• ESRI BIL/BIP/BSQ data is used to read anddisplay uncompressed, band interleaved by line,

RASTER FORMATS

band interleaved by pixel, and band sequentialimage data. Through an ASCII file that describesthe layout of the image data, black-and-white,grayscale, pseudocolor, and multiband imagedata can be displayed without translation into aproprietary format.

• ESRI ARC GRID data supports 32-bit integer and32-bit floating-point raster grids. Grids areespecially suited for both discrete andcontinuous phenomenoma and for performingspatial modeling and analysis of flows, trends,and surfaces such as hydrology.

• ESRI ARC GRID Stacks and Stack files are usedto reference multiple ESRI grids as a multibandraster data set. A stack is stored in a directorystructure similar to a grid or coverage. A stackfile is a simple text file that stores the path andname of each ESRI grid contained within it on aseparate line.

• GIF (Graphics Interchange File) is CompuServe’sstandard for defining generalized color rasterimages. This format allows high-quality, high-resolution graphics to be displayed on a varietyof graphics hardware and is intended as anexchange and display mechanism for graphicsimages.

• JFIF (JPEG File Interchange Format) applies theJPEG compression technique for storing full-color and grayscale images.

• MrSID™ (Multiresolution Seamless ImageDatabase) is a multiresolution wavelet-based imageformat with a high compression ratio. It allows fastaccess of large amounts of data at any scale.

• TIFF (Tagged Image File Format) is widely used indesktop publishing and serves as an interface toseveral scanners and graphic arts packages. TIFFsupports black-and-white, grayscale, pseudo-color,and true color images, all of which can be storedin compressed or uncompressed format.

• Microsoft Windows/IBM® OS/2® Bitmap (BMP)images are usually used to store pictures or clipart. These images can be moved between differentapplications on Windows or OS/2 platforms.


no no

yesyesyes

none single file, variousextensions

single file with .bmpextension

Supportsmultiband

yes, 1 or 3bands

yes

yes

yesyes

yes, 1 or 3bands

yesyesyesyesyes

no

Supported data types Supported compressions

wavelet

none

none

nonenone

JPEG

nonenonenonenonenone

none

File structure

single file with .sid extension

multiple files, data file with.lan extension, statistics filewith .sta extension

multiple files, header filewith .raw extension, datafile usually same name asheader without extension

single file with .jpg, .jpeg,or .jfif extension

multiple files, header filewith .ers extension, datafile is usually the samename as header filewithout extension

multiple files, data file with.gis extension, colormapwith .trl extension

nono

alwaysalwaysalwaysalways

adaptive run length compressednoneadaptive run length compressednoneadaptive run length compressednone

ARC GRID and ARC GRIDStack is a folder of files with.adf extension and colormapwith .clr extension. ARCGRID Stack file has apossible .stx extension

single file with .gif extensionno LZW

8-bit unsigned integer

8-, 16-bit unsigned integer

1-, 2-, 4-, 8-, 16-, 32-bitunsigned integer

16-, 32-bit signed integer32-, 64-bit floating point


8-, 16-bit unsigned integer32-bit unsigned integer8-, 16-bit signed integer

32-bit signed integer32-, 64-bit floating point

1-, 2-, 4-, 8-, 16-bitunsigned integer

32-bit signed integer32-bit floating point




TIFFTagged Image File Format

(GeoTIFF tags aresupported)

no

noyesyesnonono

yes

yesyesyesyesnono

none, CCITT Group 3 1-D, CCITT Group 4, PackBits, LZWnone, PackBits, LZWnone, PackBits, LZW, JPEGnonenonenonenone

single file with .tif, .tiff, or.tff extension


4-bit unsigned integer8-bit unsigned integer

16-bit unsigned integer32-bit unsigned integer

8-, 16-, 32-bit signed integer32-, 64-bit floating point

ERDAS IMAGINE yes

yesyesyes

yes

yesyesyes

none, adaptive run length compressed

none, adaptive run length compressednonenone

single file with .imgextension

1-, 2-, 4-, 8-, 16-, 32-bitunsigned integer

8-, 16-, 32-bit signed integer32-, 64-bit floating point

64-, 128-bit complex

ESRI BILBand interleaved by line

ESRI BIPBand interleaved by pixel

ESRI BSQBand sequential

yes

yes

yes

yes

yes

yes

none

none

none

multiple files, data file has.bil, .bip, or .bsq extension,header file has .hdr exten-sion, colormap file has .clr extension, statistics file




DTED Level 1 & 2Digital Terrain Elevation Data

Windows/OS/2Bitmap

BMP or MicrosoftWindows/IBM OS/2 Bitmap

or Device-IndependentBitmap (DIB)

Raster format

MrSIDMultiresolution Seamless

Image Database

ERDAS 7.5 LAN

ERDAS Raw

JFIF (JPEG)JPEG File Interchange

Format

ER Mapper

ERDAS 7.5 GIS

ESRI ARC GRID

ESRI ARC GRIDStack

ESRI ARC GRIDStack File

GIFGraphics Interchange File

16-bit signed integer

1-bit unsigned integer4-bit unsigned integer8-bit unsigned integer

Supportscolormaps

no

no

no

nono

no

yesnoyesnono

yes

yesnonononono

yes

nono

1 or 3 bands

nonenone, run length encodednone, run length encoded (single band)

yes, always3 bands

nonone multiple files with .img,.ovr, and other extensions

ADRG, ARC DigitizedRaster Graphics—image,

overview, and legend


yes, always3 bands

vector quantizationCADRGCompressed ARC

Digitized Raster Graphics

8-bit unsigned integer no single file, no standardextension

CIBControlled Image Base

8-bit unsigned integer no vector quantization no single file, no standardextension

has .stx extension



creates

is associated with

is composed of

1..*

CreateableClass

AbstractClass

InstantiableClass

multiplicity

Raster

RasterBand

Raster-Workspace

Raster-Workspace-

Factory

RasterCursor

PixelBlock

A class factory that createsRasterWorkspace objects.

A type of workspacethat contains one or

more RasterDatasets.

A collection of one or moreRasterBands storedpermanently in a particularformat.

An array of pixels andancillary data storedpermanently in a particularformat.A mechanism for enumerating

through a collection of PixelBlocksin a Raster.

A generic container for arrays ofpixels returned by Rasters.

An in-memory collection ofRasterBands intended for output.

RasterDataset

Raster-Colormap

An object that provides access to the red,green, and blue values for pixels in a

RasterBand.

Raster-Histogram

A description of the frequency of pixel values,or data distribution, in a RasterBand.

RasterTableThe collection of attributesfor a RasterBand.

Raster-Statistics

An object that provides access to the descriptive statistics fora RasterBand, which include the minimum, maximum, mean,

mode, median, and standard deviation values.

Raster objectsThis is a simplified UML diagram extracted from the ArcInfoObject Model diagram that shows the raster data access objects.

161

10 Surfacemodelingwith TINs

A surface is a continuous distribution of anattribute over a two-dimensional region. Mostoften, a surface represents the shape of theearth. But other spatial phenomena also formsurfaces, such as population density, rainfall,and atmospheric pressure gradients.

Triangulated irregular networks (TINs) are anefficient and accurate representation ofsurfaces. This chapter covers the following:

• Representing surfaces

• Structure of a TIN

• Modeling surface featuresGenua, The Gallery of Maps in the Vatican, 1636.


REPRESENTING SURFACES

Most of the geographic objects on a map lie on thesurface of the earth.

Entities such as buildings, roads, and wells areusually modeled as features—two-dimensional vectorshapes with attributes, relationships, and behavior infeature datasets inside a geodatabase.

Other entities, such as drainages, ridges, and peaks,are integral components of a surface. You canrepresent these entities as features—their shapes canbe neatly drawn on a map. But if you want toperform some form of surface analysis such ashydrography or viewshed studies, you must embedthese discrete entities within a continuousrepresentation of a surface.

The previous chapter discussed the versatile use ofraster datasets to represent a wide range ofphenomena, including surfaces. This chapter firstcompares the utility of rasters and triangulatedirregular networks (TINs) for surface modeling, andthen further examines the TIN data representation.

QUALITIES OF SURFACES

Surfaces represent a continuous field of z-valueswith an infinite number of points. Computers andthe notion of infinity are generally incompatible—some type of sampling is necessary to derive anacceptable approximation of a surface in a GIS.

ArcInfo uses two representations to model surfaces:rasters and TINs. Rasters represent a surface as aregular grid of locations with sampled orinterpolated z-values. TINs represent a surface as aset of irregularly located points that form a networkof triangles with z-values at each node.

Both raster and TIN representations have merit forsurface modeling; the context of available sourcedata, and the scope of analysis and cartography tobe supported, will guide which representation isbetter for a particular application.

The raster representation of a surface

Rasters represent surfaces as a regular grid ofuniformly spaced locations with z-values. You canestimate a surface value for any location byinterpolating z-values among immediate grid points.

The resolution of the grid—the width and height ofcells—determines the precision of the rasterrepresentation.

Rasters are the most common representation ofsurfaces because elevation data is widely available inthis form at low cost. An example of raster surfacedata are the digital elevation models (DEMs)produced by the United States Geological Survey.

Rasters support a rich set of spatial analysis such asspatial coincidence, proximity, dispersion, and least-cost paths, which can be performed rapidly.

The disadvantages of the raster representation is thatsurface discontinuities such as ridges are not wellrepresented and precise locations for features suchas peaks are lost in the grid sampling of rasters.

Rasters are appropriate for small-scale mappingapplications where positional accuracy is notparamount and where surface features do not needto be characterized exactly.

The TIN representation of a surface

TINs represent surfaces as contiguousnonoverlapping triangular faces. You can estimate asurface value for any location by simple orpolynomial interpolation of elevations in a triangle.

Because elevations are irregularly sampled in a TIN,you can apply a variable point density to areaswhere the terrain changes sharply, yielding anefficient and accurate surface model.

A TIN preserves the precise location and shape ofsurface features. Areal features such as lakes andislands are represented by a closed set of triangleedges. Linear features such as ridges are representedby a connected set of triangle edges. Mountain peaksare represented by a triangle node.

TINs support a variety of surface analyses such ascalculating elevation, slope, and aspect, performingvolume calculations, and creating profiles onalignments. The disadvantage of TINs is that they areoften not readily available and require data collection.

TINs are well suited for large-scale mappingapplications where positional accuracy and shapes ofsurface features are important.

Chapter 10 • Surface modeling with TINs • 163

Comparing rasters and TINs for representing surfaces

TIN representation of asurface

451

454

456 458

461

467

466

458

460

461

464

458

454

471

459

460461 458

463457

458

462

464

470

472

470

463

454

455

458

451 454 457 459 458

453 455 456 461 461

454 459 458 465 467

456 460 462 473 469

458 462 464 469 465

Raster representationof a surface

Surfaces can be modeledwith rasters or TINs. Eachmodel has advantages andlimitations.

Rasters are a simpler modelof a surface. Digital terraindata is widely accessible inraster format.

TINs can produce a moreaccurate representation ofsurfaces and features, butusually require a datacollection effort.

Accuracy ofsurface model

The precision of a raster surface model isdetermined by the cell dimensions. Toincrease the accuracy of a raster surfacemodel, the entire raster must beresampled at a higher resolution.

A TIN surface model has a variable pointdensity that varies on the degree ofchange of slope. To make a TIN moreaccurate, additional mass points,breaklines, and polygons can be added.

Surfaceanalysis

Spatial coincidenceProximityDispersionLeast-cost path

Elevation, slope, aspect calculationsContour derivation from surfaceVolume calculationsVertical profiles on alignmentsLine-of-sight analysis

Fidelity ofsurfacefeatures

Rasters sample the z-values of surfacefeatures on a regular grid. Features such aspeaks and ridges cannot be located to aposition more accurate than the grid resolution.

TINs are designed to capture and representsurface features such as streams, ridges, andpeaks. These features are stored with precisecoordinates, and slope discontinuities such asridges are modeled with breaklines.

Sampleapplications

Small-scale surface display and modelingModeling of pollutant dispersionIdentification of watershed basinsHydrologic analysis of flood zones

Volumetric calculations for roadway designDrainage studies for land developmentGeneration of high-quality elevation contoursPerspective displays of buildings on a landscape


STRUCTURE OF A TIN

The TIN data structure lets you accurately representany type of surface. Not only can elevations beinterpolated for any location within a TIN, butnatural features that form breaks in a surface’s slope,such as ridges and streams, can also be stored.

DEFINITION OF A TIN

4317

4327

4328

4323

4310

4313 4336

4321

4347

4350

4341

4333The TIN representation models asurface from a set of points fromwhich triangles are formed, ortriangulated.

Triangles are made from three pointsthat occur at irregular locations.

Each triangle stores topologicalinformation about its neighboringtriangles, thus forming a network.

triangulated

irregular

network

The term triangulated irregular network is a concisedescription of the characteristics of a TIN.

“Triangulated” refers to the forming of an optimizedset of triangles from a set of points. Triangles make agood representation of a local portion of a surfacebecause three points with z values uniquely define aplane in three-dimensional space.

“Irregular” identifies the key advantage of TINs forsurface modeling—points can be sampled withvariable density to model areas where change insurface relief is abrupt.

“Network” reflects the topological structure that isimplicit in a TIN. This structure enables sophisticatedsurface analysis as well as compact representation ofa surface.

Creating TINs

TINs are made from mass points, which are pointswith elevations collected from a variety of sources.TINs are commonly compiled with photogrammetricinstruments that sample elevations from pairs ofaerial photographs precisely aligned in a stereomodel. TINs are also produced from survey data,digitized contours, rasters with z-values, point sets infiles or databases, or operations on other TINs.

From these input points, a triangulation is performedon the set of points. In a TIN, the triangles are called

faces, the points become nodes to a face, and thelines of faces are called edges.

edge

node

masspoints

(x,y,z)

(x,y,z)

(x,y,z)

triangulation

(x,y,z)

(x,y,z)

(x,y,z)face

Each face in a TIN is a part of a plane in three-dimensional space. All of the faces in a TIN meettheir neighbors precisely at each node and alongeach edge. Faces cannot intersect each other.

edge

z

y

x

nodeface

Triangulation and topology

Given a set of points, many possible triangulationscan be created. ArcInfo uses an algorithm called theDelaunay triangulation to optimize how facesmodel a surface.

The basic idea of this algorithm is to create trianglesthat collectively are as close to equilateral shapes aspossible. This keeps the interpolation of elevations atnew points in closer proximity to the known inputpoints.

A triangulation can be made from an input set ofsurface features represented by points, lines, andareas. First, a triangulation is made from points.Next, lines are inserted into the triangulation andnew nodes are created wherever those lines splitfaces. Finally, areas are inserted; these can also splitor clip faces.

After triangulation is complete, the TIN stores a listof nodes for each face, and for each face, a list ofneighboring faces. This representation is similar tothe topology represented by planar topologies. Thedifference is that nodes have elevations and facesmust be triangles instead of arbitrary polygons.


Topology and triangulation

1

2

3

4

1

2

3

4

1

2

3

4

1

2

3

4

1

2

3

4

From a simple case of fourmass points, two

triangulations are possible.Which is the valid TIN?

The definition of the Delaunaytriangulation specifies that any circlearound three nodes in a triangle willnot include any other node.

This triangulation fails theDelaunay test because thecircle bounding nodes 1, 3,

and 4 includes node 2.

This triangulation satisfies theDelaunay test because a

circle around each trianglecontains no other nodes. This

is the valid triangulation.

The Delaunay Triangulation

A Delaunay triangulationuses an algorithm tooptimize the surface

representation.

A

B

C

D

E

F

G

H

1, 2, 3

2, 4, 3

4, 8, 3

1, 3, 5

1, 5, 6

3, 7, 5

3, 8, 7

5, 7, 6

–, B, D

–, C, A

–, G, B

A, F, E

D, H, –�

G, H, D

C, –, F

F, –, E

Triangle Node list Neighbors

AB

D C

GF

H

E

1

5

6

7

3

8

4

2

A TIN is a topological data structure that managesinformation about the nodes that comprise each triangle

and the neighbors to each triangle.

Triangles always have three nodes and usually have three neighboringtriangles. Triangles on the periphery of the TIN can have one or two neighbors.

Topology in a TIN


You can create a TIN by entering surface featuresthat represent elements of terrain such as pointelevations, peaks, streams, and ridges.

Point elevations are the predominant input into aTIN and form the overall shape of the surface. Theycan be input from contour lines if necessary, but it isbetter to use points collected with photogrammetricdevices because the operator can do a better job ofvisually sampling points that reflect terrain relief.

Streams, ridges, and similar surface features are thenadded to refine the surface model and sharpen thechanges in relief. These features are preserved in theTIN and increase the model accuracy.

REPRESENTING SURFACE MORPHOLOGY

This is a view of surface features as shown on manymaps. These features can be converted into a TIN.

stream course

contour

ridge

peak

Point surface features

Mass points represent points at which a z-value ismeasured. After triangulation, they are preserved asnodes with the same location and elevation.

Line surface features

Breaklines are linear features that represent naturalfeatures such as courses and ridges or man-madefeatures such as roadways. There are two kinds ofbreaklines: hard and soft.

Hard breaklines represent a slope discontinuity suchas a stream course. While the surface is alwayscontinuous, its slope may not be. Hard breaklinespreserve surface sharpness and improve the analysisand display of a TIN.

MODELING SURFACE FEATURES

Soft breaklines let you add edges to represent linefeatures, but do not represent a slope discontinuity.For example, you might want to add the imprint of aroad to your surface model, but it may notsignificantly alter the local surface slope.

Areal surface features

Polygons represent objects such as lakes or coasts.

Replace polygons assign one constant z-value to theboundary and all interior heights.

Erase polygons mark all areas within the polygon asbeing outside the zone of interpolation for the mode.Analytic operations such as volume calculations,contouring, and interpolation will ignore these areas.

Clip polygons mark all areas outside the polygon asbeing outside the zone of interpolation for themodel.

Fill polygons assign an integer attribute value to allfaces within the polygon. No height replacement,erasing, or clipping takes place.

FUNCTIONAL SURFACES

A TIN is a surface representation with a single z-value for each point. An interesting quality of theTIN data representation is that it represents points inthree-dimensional space, but the topological networkof faces is constrained to two dimensions.

For this reason, the TIN data representation issometimes said to model “two-and-a-half”dimensions. This description is not quite accurate,but illuminates the fact that surfaces have pointsmeasured in three dimensions but each point canhave only one z-value.

Consequently, TINs are an example of a single-valuefunction—given an input location, only one z-valuecan be interpolated. A minor constraint of TINs isthat they cannot model the infrequent occurrences ofnegative slope, such as overhanging cliffs and caves.Unless you are modeling climbing routes inYosemite Valley or passageways through CarlsbadCaverns, this is thankfully not a significant limitation.


Using surface features to create a TIN

BreaklineLinear surface features

Breaklines represent naturalfeatures such as ridges, or man-

made features such as roads.Breaklines can be identified as hard

or soft, which influence howelevation isolines (contours) are

drawn on a map.

When breaklines are added, newnodes are created at theintersections with the initial edges, ifthose edges represent break edges.The TIN is updated to incorporatethese new nodes.

Areal surface features

Polygons either represent arealfeatures with a constant elevation,

such as a lake or ocean, or areused to delimit the project boundary

and clip or erase portions of thetriangulation. Polygons can also be

identified as hard or soft.

The TIN has now been refined tomodel areas of constant elevationsand the boundary of interpolation.

After the TIN has reached this stage,you can inspect it and add anyadditional mass points, breaklines, orpolygons to correct and improve theTIN surface model.

Polygon

node

edge

face

hull

island

elevationpoints

stream

lake

shoreline

When you build a TIN, mass points, breaklines, and polygons are progressively added to create and refinethe surface model. At any time, you can add additional surface features to further improve the model.

TINs are created from surface features that can becategorized by dimension: points, lines, and areas.These surface features are called mass points,breaklines, and polygons, and can be input from avariety of data sources.

Surface features TIN

As surface features are input to a TIN, they become atopological network of nodes, edges, and faces. Each

node in a TIN has a z-value. A hull is a boundary ofsurface interpolation. A TIN has at least one hull on itsexterior; additional hulls can delimit islands and lakes.

Point surface features

Mass points represent locationswith known z-values. They can

represent specific features such aspeaks or the bottom of a

depression, but most are pointssampled from a surface with

measured z-values. The density ofmass points should vary by the

degree of change in the surface.

From the set of mass points, theDelaunay triangulation algorithm isapplied to create an initial TIN. ThisTIN reflects the overall shape of thesurface, but does not yet wellrepresent sharp changes in terrainsuch as streams and ridges.

Mass points

triangulation


Surface features in a sample TIN

This illustration depicts a fictional TIN designed to show some of the keytypes of surface features as they are drawn in ArcMap.

low elevation

high elevation

A TIN can be drawn withsymbols applied to elevation,aspect, slope, or othercharacteristics of a surface.This map shows elevations.

Soft breaklines or polygonrepresent surface features that donot mark a slope discontinuity, butadd edges to a TIN to better modela surface.

Hard breaklines or polygonsrepresent features that mark asignificant change in slope.

To illustrate the structureof a TIN, this mapsymbolizes nodes, edges,and faces. Normally,nodes and edges are notdisplayed on a map. In thismap, faces are drawn withinterpolated elevationranges.

node

edge

face

A lake can be modeled as an replacepolygon applying the lake elevationfor a constant height.

A TIN can model adiscontinuous surface.Each discrete land massis represented with onehull and an integratedset of nodes, edges, andfaces.

This region is steep withsharp changes in relief. Moremass points are necessary torepresent the local shape ofthis surface.A shoreline for an island or

land mass is modeled byadding an clip polygon. Thispolygon defines the hull, or

outer edge of a set of faces.The hull defines a shore or a

project boundary. In this map, a stream course is modeled with a softbreakline. For features such as stream courses and ridgelines, you can choose either soft or hard breaklinesdepending on the effect you desire when drawing a TINwith smoothed interpolation of elevations.

This area has a gentle slopewithout breaks in the slope.Few mass points arerequired here.

169

11 Findinglocations

A wealth of geographic information resides insources that are not maps. Examples areaddresses in telephone books, highwaymileposts in logs of traffic incidents, and placenames in a gazeteer. Given reference dataand methods for location, you can deducegeographic positions from location data andadd new features to your map.

This chapter discusses the following:

• Using locations

• Converting locations to map features

• Converting x,y locations

• Converting addresses

• Converting place names

• Converting postal zones

• Converting route locationsTotius Africae accuratissima tabula, Frederik de Wit, before 1710.


USING LOCATIONS

The majority of information that people trackincludes some sort of reference to a geographiclocation. For example:

• A store keeps records of customers. Thisinformation includes their addresses and the storelocation from which they have made purchases.

• A police department keep logs of crime incidents.These records reference locations oftransgressions by street intersections or addresses.

• A department of transportation keeps track ofroad maintenance through a mileposting system.

A unique quality of a geographic information systemis its ability to integrate diverse types of data into acommon geographic framework. Tying diverse datatogether gives you considerable freedom to explorethe relationships between entities such as people,highways, land, stores, and natural features.

This chapter discusses the types of locations, how todefine location methods and set up reference data toinfer geography, and the addition of point, line, orarea features to your map from location data.

TAXONOMY OF LOCATION DATA

Geographic locations come in many forms. The mostcommon locations are street addresses, but otherlocations include mileposts, names of significantplaces or features, and various types of gridlocations.

These are the types of geographic locations fromwhich you can derive features and add to yourgeodatabase.

x,y coordinates

Often, an organization will maintain tables thatcontain records with attributes for x,y coordinates.

For example, an environmental agency might createrecords of air quality readings. For each reading, acoordinate can be collected with a GPS receiver orread from a map. To study these readings in thecontext of sources of air pollution, you can convertthe collected coordinates and create points in a newfeature class. You can compare these point featureswith locations of factories to analyze the cause anddispersion of pollutants in the environment.

Street addresses

Many databases contain addresses to keep track ofcustomers and business locations. For manycountries, accurate street maps can be combinedwith methods to convert an address to a location.

For example, a corporation can maintain databasesof customer purchases that contain the customeraddress. The corporation can convert the addressesto points on a map to study the optimum geographicplacement of new stores.

Postal codes

Sometimes, databases might not contain addressesbut postal codes, such as the ZIP Code system in theUnited States. These codes can reference one toseveral thousand addresses.

For example, consumer demographic data is widelyavailable aggregated on postal zones. Someone whowants to focus a marketing campaign on a desireddemographic can perform spatial clustering analysisto optimize the cost benefit of advertising.

Place names

People find their way by making reference tolandmarks, such as government buildings, schools,and mountain peaks.

For example, a tourist bureau might developinformational kiosks that let the user select from adatabase of hotels, restaurants, and points of interest.A database with place names and positions can beused to construct maps to guide the user.

Route locations

Many of the things that are built—roads, canals,railroads—are designed with a linear measurementsystem based on distances from starting points.

For example, a department of transportation keepsextensive records of all aspects of a highway system,such as pavement quality, incidents of accidents, andsignage. Data for these entities is stored in amileposting system that references a route and lineardistance from a known point. Given route referencedata in the form of accurate street maps, any routemeasurement can be added to a map.

Chapter 11 • Finding locations • 171

Types of locations

City Hall

Many locations such asprovince, city, or important

buildings have placenames.

35 13' 45" N105 16' 21" E

An x,y location specifies apoint on the earth in a

coordinate system.

An address can beinterpolated from a streetsegment or assigned to afeature such as a building.

87501

Postal codes representareas of service for several

hundred or thousandaddresses.

Postal code

Roads, railroads, streams,and other linear systemscan have a measurementsystem applied to them.

35+00

Route location

Place name

x,y location

759 WestManhattan

Avenue

Street address


CONVERTING LOCATIONS TO MAP FEATURES

Using locators, you can convert a location to aknown position on a map.

If you can take some kind of location and figure outa precise or approximate position on your own, youcan train ArcInfo to do the same.

The general technique for adding locations to a mapis to prepare a locator, apply location data to it, anduse the point, line, or polygon features it creates ona map.

LOCATORS

A locator is a combination of reference data andlocation method. There are locators for each type oflocation.

For addresses, the reference data is a street map thatcontains street centerline segments with addressranges, street names, and related attributes. Thelocation method converts addresses to a positionalong the street segment.

For x,y locations, the reference data is the definitionof the coordinate system. The location method is thecoordinate transformation function built into ArcInfo.

For place names and postal codes, the reference datais a map with place names or codes as an attribute.The location method is a simple match between theinput place name or code and the feature thatcontains it.

For route locations, the reference data is a street mapwith routes and a measurement system. The locationmethod finds the input route, finds the line that

contains the input measure, and calculates theposition of the input measure.

Location data

The location data is a table with one or moreprescribed fields that describe the locations you wantto find on a map. Location data is applied to alocator to create points on a map.

Location feature classes

The result of the application of a locator is a newfeature class that contains points, lines, or polygons.

Depending on your application, you can create andstore the feature class or you can create temporaryfeatures each time your application reads thelocations. The location feature class has a spatialreference that is defined by the locator.

Custom locators

ArcInfo presents a number of standard locators thatare ready to use. However, some applicationsdemand sophisticated custom location methods. Anexample would be an address locator for a countrythat does not follow the postal convention of housenumbers ascending on a range along a streetsegment. For this circumstance, you might findcustom locators developed by internationaldistributors or GIS consulting firms.

The location framework in ArcInfo is open, flexible,and extensible. You can define as many customlocators as you want.

finding locations

Applying locationdata

creates new points at thecalculated positions.

These points are drawn on a map and can beused to solve a geographic problem.

if location contains prefix then look up standard prefixes apply conversion interpolate position create feature...

to a locator

which is a combination of a locationmethod and reference data


You can convert a database table containing x,yvalues into points on a map. These points arecreated in a new feature class that contains all of theattributes of the input database table.

Some common scenarios for x,y locations are pointscollected when sampling environmental data,performing field maintenance of structures, surveyingcorners of a land parcel, and tracking the motion ofa boat, airplane, or other vehicle.

The increasing use of GPS receivers makes x,ylocations more common as attributes in databasetables.

Points collected from a field survey drawn on a map

CONVERTING X,Y LOCATIONS

Input location table

The x,y values in the input location table must be intwo numeric fields. The x,y values cannot becombined into one field. The fields with x,y valuescan have any name.

The implicit coordinate system of the x,y values mustbe in a form that the ArcInfo spatial reference systemcan accept.

Locations and spatial reference

It is not necessary that the x,y values be in a spatialreference already present in your geodatabase. Forexample, x,y data is commonly collected as latitude/longitude pairs. When you create a locator for x,ylocations, you will specify the coordinate system ofthe input data and a spatial reference for the newlycreated feature class.

Therefore, records with latitude/longitude values canbe converted into a feature class with a spatialreference that can be directly overlaid on otherfeature datasets and feature classes in yourgeodatabase.

The x,y locator will check the input x,y values toensure that they fall within the expected range of thespatial reference system. For example, longitudevalues must be between −180 and 180.

table with x,y’s

field with x field with y

Specify fields with x and ycoordinate values.

Select a spatial reference.

coordinatesystem

scale

x domain

A point feature class is created withpoints for every input x,y value.

finding x,y locations

The original x,y values are preserved asfields in the point feature class.

y do

mai

n

coordinate coordinate


CONVERTING ADDRESSES

The single most common form of geographicinformation is the address. There are many moreaddresses kept in corporate and governmentdatabase tables than features in all the digital mapsever created. All businesses and governments keeptrack of people directly or indirectly through theiraddresses.

Because addresses are so prevalent in anyinformation system, an important area of work forGIS professionals is developing and updatingnational street maps with address attributes andmethods for finding addresses.

ArcInfo comes with a CD–ROM containing addressreference data for the entire United States. ArcMapincludes a set of predefined address locatorsdesigned to work with this data. Every address in theUnited States can be matched against this data—thismakes the job of finding positions for addresses inyour tables easier. Your main job will be identifyingand correcting errors in your address tables.

If you are outside the United States, check with yourinternational ESRI distributor for the availability ofnational street data with suitable address locators.

A set of addresses converted to annotated features on a map

413 State St

34 Petra Lane

221 Baker Rd

66 Webber Rd

finding addresses

On a table with addresses,specify the address fields.

Select an address locator and setspelling sensitivity.

Following the national postalconventions, positions are foundfor each address and points arecreated in a new feature class.

Standardized addresses areoptionally written to a field.

table with two address fields

table with four address fields

Address tables can contain addresses witheither two or four address fields.

table with addresses

field with house numberand street name

field with city, state, or postal code

18 Quirin Rd

221 Lena Dr

110 Michaela St

90 Francisusi Ave

43502

10010

90210

34112

Matching addresses can be ambiguousbecause of spelling errors and incompleteaddresses. After you have processed anaddress table, you will find a percentage ofpoint features for which no position wascreated. You can postprocess these missingaddresses and correct them.

address location method if US postal addresses find segment containing address interpolate number against range even addresses go on one side odd addresses go on other side...

You can matchaddresses against a

street network withaddress ranges or pointor polygon data that hasaddresses as attributes.



city field

413 Benoy Blvd

22 Lila Lane

96 Chloe Court

Pasadena

Taos

Velarde

state fieldpostal code

92373

87501

87505

CA

NM

NM


MATCHING ADDRESSES TO STREETS

The reference data for matching addresses to streetsconsists of a set of lines representing street segmentsbetween intersections. Each line has a number ofattributes such as street name and address ranges.

These are the components of addresses you mightfind in street reference data:

• Address ranges represent progressive numberingof houses along a street segment. Most localeshave address ranges on the left and right sides.

• A prefix direction denotes a direction in referenceto a local center of addressing. For example, acity may have these two addresses: 20 West MainStreet and 20 East Main Street. The prefixdirection, “West” or “East,” might be necessary tounambiguously locate an address.

• A street name is the main identifier for a streetsegment. Examples are “Main” and “Wilshire.”

• Each street has a type. Examples are “Road” and“Street.” Sometimes, street types are prefixes tostreet names, such as “Highway” and “Avenida.”

• Addresses for some cities have suffix directions,such as “NW” or “SE.”

• Street reference data can contain the left and rightpostal zones for each street segment. Thisinformation is used to validate address matching.

Street data with left and right address ranges

Street reference data can contain right and leftaddress ranges. The reference data on the U.S. streetdata CD–ROM is organized in this way.

from-rightaddress

from-leftaddress

street segment

to-leftaddress

to-rightaddress

Main Streetstreet name

250249

200201

The United States and some other countries follow apostal convention that odd addresses are located onone side of a street and even addresses on the other.

An advantage of using left and right address rangesis that positions found for addresses can be placedcorrectly on the left or right side of a street.

street reference data with right and left rangesleft-fromhouse

number

left-tohouse

number

right-fromhouse

number

right-tohouse

numberprefix

direction

prefixstreettype

streetname

streettype

suffixdirection

leftzone

rightzone

Populating these fields is mandatory.

Populating these fields is optional.

These are the fields required by the standard addresslocator for data with right and left ranges.

Street data with single address ranges

For some locales, address ranges are available forthe beginning and end of each street segment. Whenpositions are found for addresses, they are placed onthe street segment.

street reference data with one range



fromhouse

number

tohouse

numberprefix

direction

prefixstreettype

streetname

streettype

suffixdirection

These are the fields required by the standard addresslocator for data with a single address range.

Processing addresses

When the address locator processes an address table,a match score is calculated for each point. If thematch score is greater than the threshold you havedefined, a point is placed on the map. If not, a pointis still created with a null position that you canreview to correct addresses. Some causes of lowmatch scores are incomplete addresses, mispelledstreet names, or house numbers out of range.


Most of the time, you will use conventionaladdresses in your address tables; however, twoadditional techniques match addresses based onplace names and street intersections.

FINDING PLACES BY LINKING ADDRESSES

An address table can contain place names. If youhave a place name table, you can direct the addresslocator to first locate a matching place name in theplace name table, and then a matching address inthe address reference data.

Normally, this fieldcontains a house

number and a streetname, but it can alsocontain place names.

place name table816 E High

619 N Alta

227 E Palace

27182

81828

59045

State Library

La Casa Feliz Preschool

Temple Beth Shalom

table with addressesState Library

La Casa Feliz Preschool

Temple Beth Shalom

31415

92653

58979

The address table is linked to theplace name table by using the place

name as a foreign key.

From the record in theplace name table that

matches a placename, an address

match is made againstyour address

reference data.

This is a two-level matching process.

FINDING STREET INTERSECTIONS

The intersection of two streets marks a point on amap. You can add street intersections to an addresstable and find the position of the crossing.

Using the same address reference data as they wouldfor address matching, the standard address locatorscan take combinations of streets and locate theirintersections on a map.

Street intersections are specified by the use of aconnector such as “AND”, “&”, “/”, or other. You candefine a set of valid connectors for your addresstables, but you should take care that they are nototherwise used as a component of a street name.


A field with house numberand street name can also

contain street intersections.

St Francis and Cerrillos

Wood / Gormley

Paseo de Peralta & Palace

US 285 at Nambe Road

The intersection of two streetsis specified by a connector

such as “and” and “&”.

This technique creates new points for intersections.

FINDING BUILDING ADDRESSES

Another technique is to match addresses to referencedata that contains points or polygons representingbuildings, and that has address fields.



109 Montezuma

111 Montezuma

115 Montezuma

121 Montezuma

115 Montezuma

121 Montezuma

111 Montezuma

109 Montezuma

This address reference data is a point or polygonfeature class that contains several address fields. Justas with other addresses, you will specify a matchscore to set a threshold for qualifying matchedaddresses.

The standardaddress locator

uses these fields inthe building

reference data.

building reference data



housenumber

prefixdirection

prefixstreettype

streetname

streettype

suffixdirection zone

Points are created for matches on a point referencefeature class. Polygons are created for matches on apolygon reference feature class.


CONVERTING PLACE NAMES

On the previous page, you saw how place namescan be located through a two-level matchingprocess—first to a place name table and then toaddress reference data based on street segments withaddress ranges.

Instead of a street network with address ranges, youmight instead have reference data with points orpolygons and place names as attributes. Forexample, you might have a nationwide map of allthe counties or local jurisdictions. This reference datawould not naturally contain addresses; those entitiesare too large for an address to be meaningful.

A place name can have one part or two parts. Forexample, you could divide county names into twoparts: the name and the county type. Examples ofcounty types in the United States are “County,”“Parish,” and “Borough.”

If the place name table has one-part names, theplace name reference data must have one field forplace names. Similarly, if the place name table has

two fields for place names, the place name referencetable must also have two fields.

If you match a place name table against a pointfeature class, the new feature class will containpoints. If you match a place name table against apolygon feature class, the new feature class willcontain polygons.

A set of place names matched to a set of polygons representingcounties

Specify a table withplace names.

Select a place name locator that specifies aplace name field and point or polygon

reference data.

A feature class is createdwith polygons or points forevery matched place name.

If the reference data containspolygons, the features

created are polygons thatmatch the place name.

If the reference data containspoints, the features created are

points that match the place name.

Place names found on polygon data

Place names found on point data

finding place names

table with place namesJefferson

Boudin

Matanuska

County

Parish

Borough


CONVERTING POSTAL ZONES

If you are matching postal codes in the UnitedStates, you can use a ZIP Code locator.

A ZIP Code has two parts: a five-digit prefix thatidentifies an area served by a post office that mayhave several thousand addresses, and a four-digitextension that identifies individual addresses.

You can match either type of ZIP Code. For ZIP+4codes, you can choose from three formats: as onenumber, such as “875051357”; with a space, such as“87505 1357”; or with a hyphen, such as “87505-1357.”

Using ZIP+4 codes will give you the most accurateposition because you will be matching to referencedata with a greater number of points. Using ZIP+4ranges will give you less accuracy because you willbe matching to reference data with fewer points.Matching five-digit ZIP Codes will be the leastaccurate because you will match to the centroid ofthe five-digit ZIP area.

A set of ZIP Codes matched to the centroids of polygonsrepresenting ZIP Code areas.

Specify a table withZIP Codes.

Select a ZIP Code locator that specifiesU.S. ZIP Code reference data

A point feature class is created withpoints for every input ZIP Code.

8750187505

87508

All ZIP Code tables use the same U.S. addressreference data, but ZIP+4 codes produce a more

precise location.

table with ZIP Codes

field with ZIP Code

field contains five-digit ZIP Code

field containsZIP+4 code

Five-digit ZIP Codematch to centroids ofZIP Code polygons.

ZIP+4 code rangesmatch to more precise

locations.

ZIP+4 codes match tothe most precise

locations.

87505-008087505-0040

87504-0210

finding ZIP Codes

table with ZIP Codes

field with five-digit ZIP Codefield with first ZIP+4 range

field with second ZIP+4 range


CONVERTING ROUTE LOCATIONS

Certain locations are measures along linear systems.For example, road maintenance data kept by adepartment of transportation referencesmeasurements of locations along routes.

A route is a collection of polylines that have acommon identifier and contain measures. (To reviewpolylines and measures, read chapter 6, “The shapeof features.”)

A route location can represent either a point along aroute or a segment along a route between twopoints.

A route location representing a point on a routecontains a route ID and a single measure value. Theroute ID specifies the polylines with measures tosearch for. The measure value identifies the pointthat is interpolated on measures of one polyline.

A route location representing a segment along aroute has a route ID and two measure values. Thesegment is interpolated between the two measurevalues, and a line feature class contains the resultingpolylines.

The route reference data is a line feature class withroute IDs assigned to polylines and measuresestablished to each point in the polyline.

Points representing traffic accident locations.

Specify a table with route IDsand one or two measures.

Specify a route locator withroute reference data.

A point feature class is created with pointsfor each found single measure, or...

...a line feature class is createdwith polylines for the interval

between two measures.

table with routes

field withroute ID

one or two fields withroute measures

finding route locations

Route reference data is set up bydefining route IDs for polylines withmeasures for each point in a polyline.

polyline with two parts

100

111123 135 147

160

172 180

189

321330 345 362measures

route ID: 341

181

12 Geodatabasedesign guide

The geographic data model, implemented in ageodatabase design, is the foundation for allactivities with a GIS—creating expressivemaps, retrieving information, and performingspatial analysis. Designing a geodatabase tomeet these goals is a deliberate process.


• Purpose and goals of design

• Overview of design steps

• Step 1: Model the user’s view

• Step 2: Define entities and relationships

• Step 3: Identify representation of entities

• Step 4: Match to ArcInfo data model

• Step 5: Organize into geographic datasetsGeneralis totius Imperii Russorum novissima tabula, Johann BaptistHomann, before 1724.


A geographic information system has the potential tohelp your organization accomplish a myriad of tasks,from daily operations to long-term planning.Effectively implementing your GIS allows you torealize this potential, offering efficient ways toperform functions, store and share data betweenorganizational units, and integrate with othertechnologies.

This chapter reviews the general steps to achieveyour own geodatabase design.

THE NEED FOR DESIGN

What makes GIS implementation effective is a gooddatabase design. And what makes a database designgood is asking the right questions:

• How can GIS technology be implemented tostreamline existing functions, or change the way aparticular goal is achieved?

• What data will benefit the organization most?

• What data can be stored?

• Who is, or should be, responsible for maintainingthe database?

How you answer these questions will deepen yourunderstanding of GIS technology, as well as providenew insight into your organization and its functions.

Design for GIS implementation is like any otherdesign. It starts with understanding goals andprogresses through increasing levels of detail asinformation is gathered and you approachimplementation.

An example of this is a transportation model thatbegins by studying existing traffic flow and applyingpatterns of population growth. This model would bedeveloped in a series of steps with progressivelymore detail. The preliminary proposal and budgetevolves to detailed engineering drawings.

Because it is time consuming and produces no end-use applications, the design process often receiveslittle attention, if any. There are risks associated withavoiding design. If you do not go through the designprocess, you risk having a poorly constructeddatabase that does not meet your requirements, now

or in the future. This can result from duplicate,missing, or unnecessary data; inappropriaterepresentation of data; or lack of proper datamanagement techniques.

This section focuses on database design; however,you will quickly realize that the database and theapplications it serves cannot be treated entirelyindependently. As you progress through thedatabase design, you should also define theapplications that will create, use, and manage thedata.

OBJECTIVES OF DESIGN

Design is the process in which goals are defined,design alternatives are identified, analyzed, andevaluated, and an implementation plan is agreedupon. At the highest level, the design provides apicture of where you are, where you are going, andhow to get from one place to the other. As youprogress through the design, you increase detail,adding data definitions and assigning the appropriateArcInfo spatial data structures.

A database design provides a comprehensivearchitecture for the database. The design allows youto view the database in its entirety and evaluate howthe various aspects of it need to interact. Expendingtime and money to identify and resolve design issuesearly saves having to expend greater resources latertrying to solve what may well have becomeinsurmountable problems.

A good design results in a well-constructed,functionally and operationally efficient database that:

• Satisfies objectives and supports organizationalrequirements.

• Contains all necessary data but no redundant data(unless explicitly planned and properlydocumented).

• Organizes data so that different users access thesame data.

• Accommodates different views of the data.

• Distinguishes applications that maintain data fromthose that use it.

PURPOSE AND GOALS OF DESIGN

Chapter 12 • Geodatabase design guide • 183

• Appropriately represents, codes, and organizesgeographic features.

Design is time consuming and intensive, but you willbenefit from:

• Increased flexibility of data retrieval and analysis.

• Increased likelihood of users developingapplications.

• Distributed cost of data capture, storage, and use.

• Facilitated data that supports many different uses.

• Maintained data that supports many differentusers.

• Extensibility that readily accommodates futurefunctionality.

• Minimized data redundancy.

DESIGN GUIDELINES

The design process can be quite substantial. Here issome advice to ease the process and help ensuresuccess:

• Involve users. By contributing, they will gain asense of ownership and you will gain invaluableknowledge for your geodatabase design.

• Take it one step at a time. It is not necessary tocreate a complete detailed design all at once;design is an interactive and iterative process. Youcan progress in stages as appropriate for theneeds of your organization.

• Build a team. A wide range of information, skills,and decision making is required during thisprocess. At different stages, your team willcomprise various experts throughout yourenterprise.

• Be creative. The initiation of a new project is agood opportunity to survey new technology andprocesses. There is considerable potential toenhance how GIS serves your organization’s goalsand objectives.

• Create deliverables. It is best to divide a largeproject into discrete and identifiable units ofwork. Project milestones should be defined to be

no less frequent than two months or so. This willkeep your project focused and earn managementsupport.

• Keep organizational goals and objectives in focus.It is essential that the design and implementationprocess always be focused on the realrequirements of your organization and itscustomers.

• Do not add detail prematurely. Add detail at theappropriate step. For example, do not try todefine all of the validation rules for feature classesbefore geodatabases are constructed. Selectivelyintroduce implementation details throughout theproject so that the team can progress to the nextstep.

• Document carefully. The more complex theenvironment, the greater the benefit fromdocumenting your design. The use of business-diagramming software is especially useful tocommunicate your design.

• Be flexible. The initial design will not be the finaldesign as implemented. The design will evolve asyour organization changes, new technology isintroduced, and people become more adept withthe technology.

• Plan from your model. Create an implementationplan that addresses your organization’s keypriorities in a manageable fashion. If you need tocreate new datasets, build the data managementapplications first.


This database design process is presented as a seriesof steps. While by-products of your design activitiescan include identifying applications, identifyingeducational and training requirements, and settingstandards for data collection and maintenance, onlydatabase design is covered in this section.

The process is not meant to present a formalmethodology that is beyond the scope of thissection. The intent is more to guide you through adesign if you do not already use a formalmethodology.

The steps are:

• Model the user’s view.

• Define entities and their relationships.

• Identify representation of entities.

• Match to the geodatabase data model.

• Organize into geographic datasets.

The first three steps develop the conceptual model,classifying features based on an understanding of thedata required to support the organization’s functions,and deciding their spatial representation (point, line,area, image, surface, or nongeographic). The last twosteps develop the logical data model, matching theconceptual models to ArcInfo geographic datasets.

OVERVIEW OF DESIGN STEPS


123Select geographic

representation.

Define objects andrelationships.

Model the user’sview of data.

4Match to geodatabaseelements.

5Organize geodatabasestructure.

Steps in building a geodatabase

Identify and describe objects.Specify relationships between objects.Document model in diagram.

Building

Land parcel

Person

Represent features with points, lines, and areas.Characterize continuous phenomena with rasters.Model surfaces with TINs or rasters.

Determine geometry type of discrete features.Specify relationships between features.Implement attribute types for objects.

Organize systems of features.Define topological associations.Assign coordinate systems.Define relationships and rules.

Geodatabase

Feature dataset

Geometric network

Feature class

Identify organizational functions.Determine data needed to support functions.Organize data into logical groupings.


The objective of this step is to ensure a commonunderstanding between the design team and thosewho have a vested interest in the implementation ofyour GIS.

In this step, you:

• Identify the functions that support theorganization’s goals and objectives.

• Identify the data required to support thefunctions.

• Organize the data into logical sets of features.

• Define an initial implementation plan.

• Identify organizational functions.

An anticipated benefit of your GIS implementation isimprovement in the way your organization conductsits business.

IDENTIFY ORGANIZATIONAL FUNCTIONS

An organization performs business functions thataddress its goals and objectives. These functions arethe starting point for your database design. Youwork with business functions rather thanorganizational units, because functions are morestable than organization. That is, a functionperformed by one department today may beperformed by another department next year.

To begin, identify each of the functions within scopeof your project. For each function identified, providea general description of the activities that fall withinthat function. Activities may include managing theland development approval process, controlling landuse, and developing agreements for infrastructureconstruction by a developer.

Apart from the users themselves, documents andmaps serve as good information sources. Look forgeneral publications, strategic plans, and informationsystems plans.

LOCATE DATA SOURCES

Once the functions are compiled, identify the datathat supports them. Determine whether the function“creates” or simply “uses” the data.

In general, you work with two kinds of data: thedata of interest in your field and background data.

Naturally, the data of interest will be modeled inmore detail.

You can analyze each function’s scope by examininginteractions with other functions and externalplayers. Most often, data that flows out from thefunction has been created by the function. Thisindicates that it is responsible for the definition,collection, storage, and distribution of that data.

Data that flows into the function is generally theresponsibility of another function, though datareceived from an external organization may bestored and managed internally. Exchanges are inmany forms, including data, guidelines, requests, andresponses.

The question to answer at this stage is, “Who orwhat does this function interact with and what is thenature of that interaction?”

In relating data to the functions that create and storethem, you may discover synonyms, polynyms, andfunctions that duplicate the capture and storage ofdata. These situations should be resolvedimmediately or at least kept in a log for futureresolution.

This should be an interactive step with those whoperform the function—after all, they know what theydo, who they interact with, and what information isexchanged. After documenting the required data, besure to give them an opportunity to validate thediagram and any supporting text.

ORGANIZE DATA INTO LOGICAL GROUPINGS

Make a top-level grouping of all the data you expectto interact with in your GIS. These groupingsrepresent systems such as “water utility,” “landrecords,” “streets,” and “terrain.”

Each of these groupings is operated by a function toeither receive or transmit information. An example isthat a surface model with rainfall amounts transmitshydrological data to a stream network.

Each of these groupings should have a commoncoordinate system, topological type (network, planar,or none), and generally interact with each other.

STEP 1: MODEL THE USER’S VIEW


1 Model the user’s view of data

Identify organizational functions

Determine data needed to support functions

Organize data into logical groupings

applicationreview

accessor parcel

parcel ownership

subdivision

master plan

land-use zoneapplication

subdivision

street

transportation plan

Land developmentmanagement

receive applicationverify acceptance conditions

distribute for review

compile reviewscommunicate with applicant

update application dataupdate subdivision

Land-useplanning

Landdeveloper

Roadwaymanagement

Transportationplanner

Assessment

Applicationreview

Water utility

Land records

Terrain

Streets

The geodatabase design will be influenced by thestructure of your organization. Distinct departments mayhave responsibility for different segments of thegeographic data.

At a basic level, you begin by identifying the providersand consumers of geographic information. The key dataflows are modeled. This is the starting point foridentifying logical groupings of data.

Parcel

Easement

Parcel description

Parcel photograph

Owner

Address

Land records

Subdivision plats

Engineering records

Land title

Historic archive

Land assessment

Phone database

Types of data Data source

For each function, identify all of thetypes of data that are necessary to

fulfill this group’s requirement todeliver information.

For each data type, identify thelikely source of data. A part of theproject plan must include anestimate for cost of data capture,processing, and validation.

From an inventory of all the typesof geographic data that anorganization maintains, identify amodest set of groupings thatcomprise all of your geographicdata systems.


The previous step determined the broadclassification of functions, data, and the relationshipsbetween them. In this step, you examine the dataclassification more closely, identifying distinguishableobjects, called entities, that have a common set ofproperties.

You will:

• Identify and describe entities.

• Identify and describe the relationships amongthese entities.

• Document the entities and relationships with UMLdiagrams.

It is recommended that you document this designusing business graphics software such as Visio®. Onthis diagram, you would have boxes for entities andlines for relationships.

This step is significant because it adds detail to theuser’s view of the data they work with. It is mostimportant that users be involved in the definitionand validation of the models produced in this step.

You will deal with a lot of data during this step. Topartition the task into manageable units, focus onone function at a time. This will guide which data tofocus on. It may take several iterations to clarify thedefinitions of entities and their relationships.

Articulating entities and relationships

Identify entities and relationships by interpretingstatements. Nouns tend to be entities while verbsdefine relationships between entities.

• A valve controls the flow of gas. This statementdescribes an entity.

• A gas device connects to one or more gas lines.This statement describes a structural relationshipbetween entities.

• A gas system is composed of gas devices and gaslines. This statement describes the aggregation ofentities to make a new, more complex entity.

• A gas main is a type of gas line. This statementdescribes a subclassification of entities.

STEP 2: DEFINE ENTITIES AND RELATIONSHIPS

Be aware of verbs masquerading as nouns(connection, description, identification, andaggregation). These tend to obscure therelationships.

Documenting entities and relationships

A concise and clear way to document this stage ofthe design is to create simple UML diagrams. Reviewthe end of chapter 1, “Object modeling andgeodatabases,” for a quick primer on UML notation.

UML is used throughout this book and otherdiagrams available with ArcInfo to document theArcInfo system architecture. UML is also appropriatefor documenting your data model.

This is what a portion of your diagram might looklike at this stage:

MainLine

Pressurized-Main

LateralLine

GravityMain

WaterLine

NetworkLine

LineProtector * 0..1

This diagram states the following:

• A water line is a type of network line.

• A main line and a lateral line make up a type ofwater line.

• A main line can be associated with zero to manyline protectors. A line protector can be associatedwith zero or one main line.

• A pressurized main and gravity main are types ofmain lines.


2 Define objects and relationshipsBuilding

Landparcel

Person

Identify and describe objects

Document model in diagram

Pump

Meter

Meter box

Valve

Water main

Treatment plant

entity

Parcel

Easement

Parcel description

Parcel photograph

Owner

Address

Street

Bridge

Name

Traffic light

Bus route

Bus stop

Historic monument

Fence

Vegetation cover

Place names

River valley

Satellite image

–

–

Meter

–

–

–

related to

–

–

Parcel

–

Parcel

–

–

–

Street

–

–

–

–

–

–

–

–

–

Water utility

Land records

Environment

Streets

identify entities and theirrelationships

Specify relationships between objects

I own this property.

The land title lists me as owner.

A meter box iscomposed of meters.

A street name has a relationshipwith a street feature.

A valve controls the flow ofwater.

A water device connects toone or more water mains.

A water system is composedof devices and water lines.

A water main is a type ofwater line.

Form sentences that state theentities and their behavior. Thenouns are entities and theverbs are relationships.

This step can be done bywriting a progressive series ofstatements starting with “awater system is composed ofdevices and water lines.” Eachstatement should be simple andaccurate.

Many entities have closerelationships with otherentities. Relationships guideyour geodatabase design.

Once you have collected yourlist of entities and relationships,it is a good practice to create adata model diagram.

Using business graphicssoftware, start by making boxesfor entities and lines with arrows

for relationships. This diagramwill facilitate discussion with

domain experts and advance therefinement of the model.

�Winding

Way


In this step, you classify entities by the type ofrepresentation. Some entities will have a geometricrepresentation with corresponding attributes; theseare classified by their geometric characteristics. Otherentities will be represented by alphanumericinformation only, still others by images,photographs, or drawings.

Consider whether:

• The feature might be represented on a map.

• The shape of the feature might be significant inperforming geographic analysis.

• The feature is data that can be accessed andvisualized through its relationship with anotherfeature (for example, ownership informationabout a parcel can be accessed by selecting aparcel).

• The feature will have different representations atdifferent map scales.

• Textual attributes of the feature will be displayedon the screen or map products.

The following terms are provided for assigning atype. The information developed during this stepshould be cataloged as part of the feature’s datadictionary entry.

• Point—illustrates the location of a feature whoseshape is too small to be defined as an area on amap of a given scale.

• Line—illustrates the location of a feature whoseshape is too narrow to be defined as an area on amap of a given scale.

• Area—illustrates the location and polygonal shapeof a feature on a map of a given scale.

• Surface—illustrates the shape of a feature as in an“area,” but also includes shape resulting fromchanges in elevation.

• Raster—represents an area using rectangular cells(satellite image, aerial photograph, continuousdata) and can be used for analysis.

• Image, photo, drawing—each represents a digitalpicture and cannot be used for analysis.

STEP 3: IDENTIFY REPRESENTATION OF ENTITIES

• Object—identifies a feature for which no point,line, or area is required, and for which there is nogeometric or graphic representation.

If features could be represented in two formsdepending on scale, identify both possibilities in thedata dictionary, and use the more complexrepresentation for consideration in the remainder ofthe analysis.


3 Select geographic representation

Pump

Meter

Meter box

Valve

Water main

Treatment plant

entity

Parcel

Easement

Parcel description

Parcel photograph

Owner

Address

Street

Bridge

Name

Traffic light

Bus route

Bus stop

Historic monument

Fence

Vegetation cover

Place names

River valley

Satellite image

point

point

point

point

line

point

spatial type

area

line

text

image

object

location

line

point

text

point

line

point

point

line

area

text

surface

image

–

–

Meter

–

–

–

related to

–

–

Parcel

–

Parcel

–

–

–

Street

–

–

–

–

–

–

–

–

–

Water utility

Land records

Environment

Streets

set spatial representationas vector, raster, and TIN Represent discrete features with points, lines, areas

You can model the richestexpression of features with thevector types. These entities arewell defined on a map and arepermanent.

Characterize continuous phenomena with images

Images have versatileapplication in a GIS. You wouldspecify images for aerial orsatellite photographs,photographs of facilities, andany scanned documents.

Model terrain with surfacesWhen you model a continuousphenomenon that has a zvalue, specify surface. (Later,you will decide whether TIN orraster is better for the surface.)

point

line

area

annotation

object

an entity too small to map with a line or area

a long entity too narrow to map with an area

an entity with length and width at the map scale

a descriptive label on an entity

a nongeographic entity, such as an owner

image a file that contains a continuous valued map,

aerial photograph, copy of a plat, or picture of a

building

surface a system of points or locations with elevation

values that form a mesh for a mathetical

approximation of the shape of the earth


The objective of this step is to determine how data isto be represented in ArcInfo. For each of the spatialtypes identified in the previous step, you now assigna corresponding ArcInfo representation.

The focus now turns from understanding the userrequirements to developing an efficient and effectivedatabase schema. It is important that the team havemembers who understand the geodatabase datamodel and analysis capabilities as well as other datamanagement technologies to be used for yourdatabase.

In this step, you:

• Determine the appropriate geodatabaserepresentation for entities.

• Ensure that complex feature classes aresupported.

DETERMINE GEODATABASE REPRESENTATION

ArcInfo lets you store discrete entities as simplefeatures, complex features, and objects.

If the spatial type is point:

• For an unconnected point, such as a historicalmonument, enter a point feature.

• For a connected point, such as an intersectionconnected to street segments, enter a simplejunction feature.

• For a connected point that has an internaltopology, such as a treatment plant, enter acomplex junction.

If the spatial type is line:

• For a stand-alone line, such as a fence, enter aline feature.

• For a linear feature that participates in a systemsuch as a road network, enter a simple edgefeature.

• For a linear feature with connected sections, suchas a section of utility line, enter a complex edgefeature.

STEP 4: MATCH TO GEODATABASE DATA MODEL

If the spatial type is area:

• For a stand-alone area, such as a park, enter apolygon feature.

• For space-filling areas, such as vegetation cover,enter a polygon feature (later assigned to a planartopology).

If the spatial type is image (photograph, scannedmap, satellite image, or other), enter a raster.

If the spatial type is surface:

• For surfaces in which terrain detail is important,enter a TIN.

• For surfaces covering large areas and to utilizeexisting digital elevation models, enter a raster.

If the spatial type is an object, enter an object. Theseare entities that do not have direct geographicrepresentation, but are related to geographicfeatures.


4 Match to geodatabase elements

Pump

Meter

Meter box

Valve

Water main

Treatment plant

entity

Parcel

Easement

Parcel description

Parcel photograph

Owner

Address

Street

Bridge

Name

Traffic light

Bus route

Bus stop

Historic monument

Fence

Vegetation cover

Place names

River valley

Satellite image

object

point feature

point feature

simple junction

complex edge

complex junction

ArcInfo type

polygon feature

line feature

annotation feature

raster

object

address

line feature

point feature

annotation feature

point feature

line feature

point feature

point feature

line feature

polygon feature

annotation feature

TIN

raster

point

point

point

point

line

point

spatial type

area

line

text

image

object

location

line

point

text

point

line

point

point

line

area

text

surface

image

Water utility

Land records

Environment

Streets

–�

–�

Meter

–�

–�

–�

related to

–�

–�

Parcel

–�

Parcel

–�

–�

–�

Street

–�

–�

–�

–�

–�

–�

–�

–�

–�

apply feature geometryand topology

real values

Implement attribute types for objects

short integer

long integer

float

double

text

date

objectID

blob

short integer

long integer

float

double

text

date

objectID

BLOB

2.310.0

-4.7 8.63

time

identifiers

multimedia

whole numbers-14 56

64143

2 September 1999, 8:20

flightOver.mov

239648547593

descriptionspurple mountains

Each entity can have many attributes. These arethe attribute types.

Determine feature and geometry type

Row

Junction-Feature

Edge-Feature

Network-Feature

Feature

Complex-Edge-

Feature

Simple-Edge-

Feature



For nongeographicobjects, select row.

For simple geographicobjects, select feature.

For features in a network,select simple or complexedge or junction feature.

Specify topological graphs

For linear systems, such astransportation or utility, selectgeometric network.

A geometric network has custombehavior built in to make the editingof networks easy.

For systems of land or jurisdictions,a planar topology manages theshared geometry of a set offeatures.

A planar topology enforces that nofeature can cross another without anintersection.


The objective of this step is to identify and name thegeographic datasets that will contain the variousentities, and in the case of the coverage dataset, toorganize entities into coverages.

In this step, you will

• Assign entities to feature classes and subtypes.

• Group related sets of features into geometricnetworks or planar topologies.

• Organize feature classes and datasets intogeodatabases.

Grouping feature classes

In the previous step, you assigned feature types andattributes to entities. Now, you will define thestructure of feature classes with subtypes andwhether they stand as separate feature classes or arecontained within a feature dataset.

Your first consideration is whether an entity shouldbe mapped to a subtype or an entire feature class.Your preference should be to consolidate relatedentities as subtypes within a feature class, becausefewer feature classes will yield better-performinggeodatabases. Here are the circumstances when it isnecessary to create new feature classes instead:

• When each group of related features requiresdistinct custom behavior.

• When the set of feature attributes is substantiallydifferent. (All features in a feature class have thesame set of attributes.)

• When you require distinct access privileges foreach group of features.

• When some features are to be accessed throughversions and some are not.

Defining topological roles for feature classes

You have defined the feature types for entities.

If the feature type is simple edge, simple junction,complex edge, or complex junction, then itparticipates within a geometric network. All thefeature classes for a geometric network must beplaced within a feature dataset. This enforces thatthey share a common spatial reference.

STEP 5: ORGANIZE INTO GEOGRAPHIC DATASETS

If the entity feature type is line or polygon and theentity is either meant to cover a complete area, suchas land parcels, or if you wish to enforce thatcrossing features have intersections, then place thosefeatures within a common feature dataset. In theArcMap Editor, you can perform topological editingon these feature classes. This assemblage is called aplanar topology.

For entities with simple features, you can also placethem within a feature dataset, which also serves as acontainer for you to arbitrarily group feature classesthat are similar.

Gathering datasets and feature classes

Once you have defined your set of feature classesand their topological associations, it is time to groupthem into geodatabases.

These are some considerations for grouping featureclasses and feature datasets into distinctgeodatabases:

• If you are working in a large organization,different departments have responsibility forvarious datasets. Geodatabases can be deployedto follow your organizational structure.

• You have the freedom to use any number ofcommercial relationship databases, but each mustbe served through a separate geodatabase.

• If you are working with personal geodatabases,practical size limits may require thematic or spatialpartitioning of geodatabases.


5 Organize geodatabase structureGeodatabase

Feature dataset

Geometric network

Feature class

Pump

Meter

Meter box

Valve

Water main

Treatment plant

entity

Parcel

Easement

Parcel description

Parcel photograph

Owner

Address

Street

Bridge

Name

Traffic light

Bus route

Bus stop

Historic monument

Fence

Vegetation cover

Place names

River valley

Satellite image

object

point feature

point feature

simple junction

complex edge

complex junction

ArcInfo type

polygon feature

line feature

annotation feature

raster

object

address

polyline feature

point feature

annotation feature

point feature

line feature

point feature

point feature

line feature

polygon feature

annotation feature

TIN

raster

point

point

point

point

line

point

spatial type

area

line

text

image

object

location

line

point

text

point

line

point

point

line

area

text

surface

image

Water utility

Land records

Environment

Streets

–�

–�

Meter

–�

–�

–�

related to

–�

–�

Parcel

–�

Parcel

–�

–�

–�

Street

–�

–�

–�

–�

–�

–�

–�

–�

–�

Land base

Land parcels

Lot images

US Postal

geodatabase

polygon feature class

annotation feature class

object class

raster datasetand rasters

locator and address

Image

Address

Description

Easement

Parcel

Owner

feature dataset

line feature class

relationship class Ownership

Subdivisionplanar topology

Streets


point feature class

line feaure class

Name

Bridge

Street

Traffic light

feature dataset

line feature class

point feature class

Bus route

line feature class

Environmentpoint feature class


annotation feature class

Vegetation

Fence

Monument

Names

feature dataset

line feature class

TIN dataset

Landsatraster datasetand rasters Images

Bus stop

Valley

WaterSystem

WaterFeatures

Valve

MeterBox

Meter

WaterMain

Pump

TreatmentPlant

geodatabase

point feature class

simple junction feature class

complex edge feature class

feature dataset

point feature class

object class

complex junction feature class

WaterNetworkgeometric network

Index

A

abstract class 19address tables 174addresses 8, 174

components of 175processing 175

aerial photography 25, 148aggregation 20angles 105annotation 8, 26, 66, 95ArcCatalog 10, 12, 14, 19, 20,

21, 62, 121, 123ArcMap 14, 19, 21, 27, 57,

121, 142, 174ArcObjects 14, 16, 19arcs. See linesArcSDE 12, 14, 62, 121ArcView GIS 68areas 190areas of exclusion 51, 53aspect 38, 43, 53, 151associations 19attribute domains

8, 77, 78, 88, 90coded values 90default values 90range 90split and merge policies 90

attribute indexes 82attributes 17, 26, 76, 78, 86

BLOBs 86coded values 26, 86dates 86descriptive strings 26, 86numeric values 26, 86object identifiers 26, 86

B

basemaps 150Bézier curves 84, 103Boolean operators 110breaklines 51, 53, 166

C

CAD data model 4CAD files 68Cartesian coordinate system 76cells 8, 25, 38, 149

attributes 152circular arcs 84, 103classes 17

types of 19classifications 33

defined interval 36equal interval 36natural breaks 36quantile 36standard deviation 37

color ramps 34, 39columns 14, 82complex edge features 134, 136complex junction features

134, 137composition 20computer-aided software

engineering 20, 96connections

database 62folder 62

connectivity table 132constructors

angle 109circular arc 107curve 108line 107multipoint 106path 109point 105

contours 52coordinate geometry 105coverage data model

4, 21, 84, 102coverages 66createable class 19

D

data frames 27data representations

locations 8raster 8, 51, 60

triangulated 8, 51, 60vector 8, 51, 60

default version 121Delaunay triangulation 53, 164digital elevation models

52, 150, 151dynamic link library 96

E

edges 128TIN 164

elevation 42, 53elliptical arcs 84, 103encapsulation 10entities 188

classifying 190envelopes 102, 103environmental analysis 150

F

faces 41, 164feature classes

51, 64, 80, 102, 194and geometric networks 130annotation 95location 172network 130subtypes 76, 78, 80, 88, 194

feature datasets8, 51, 64, 80, 194

features 96advantages of 102cartographic display 6custom 10, 77, 96, 138geometry 76, 84, 102interactive analysis 7network 57relationships among 6simple 57topological 57topological roles 194

fields. See attributes


G

geodatabase data access objects10, 14

geodatabase data model5, 7, 78, 102, 184

benefits of 7geodatabases 64, 194

adding intelligence 78and versioning 118designing 14, 16, 80, 182enhancing relational databases

47extending databases 12multiuser 12personal 12, 62representations of data 8steps to building 18

geographic features25, 51, 56, 192

geographic information systemsdefined 46diverse applications 48implementing 182managing data 62purpose of 2

geometric networks8, 57, 77, 80, 194

barriers 142benefits 128disabled features 140flow direction 139indeterminate flow 140–141NetFlags 142sources and sinks 139–140uninitialized flow 141weights 141, 143

georeferencing 156Global Positioning System 47

H

hillshading 42, 151hydrological analysis 150

I

image. See rasterimage data 149INFO 66inheritance 10, 20instantiable class 19instantiation 20interval data 152

J

junctions 128

L

labels 30latitude/longitude values 173layers 27, 28, 70

feature 29, 32raster 29, 38TIN 29, 41

linear measurement systems104, 179

lines 8, 25, 53, 54–55, 56, 66,103, 190, 192

locations 170place name 170, 176, 177postal code 170route 170, 179street address 170street intersection 176x,y 170, 173

locators 8defined 172

logical data model 16, 184logical networks 130long transactions 119

M

m values 104map document 70map elements 27map scale 27, 28map surrounds 27maps 24, 27mass points 53, 166

measures 179Microsoft Access 12, 62Microsoft Component Object

Model 21, 96multiplicities 20multipoints 84, 102

N

network analysis 142network elements 130network features 130networks

transportation 139utility 139

nodes 66TIN 164

nominal data 152nonplanarity 130normal 41normalization 37notifications 94null geometries 84null values 37

O

object classes 64, 96object orientation 10objects 17, 78, 96, 190ODBC 62Open GIS Consortium 12, 14optimistic concurrency 119ordinal data 152

P

paths 103physical database model 17picture data 38, 148, 149planar topologies

8, 57, 77, 80, 194point IDs 104points 8, 25, 53, 54, 56, 66,

76, 84, 102, 190, 192polygons 8, 25, 53, 55, 56, 66,

76, 84, 102, 103, 192in a TIN 166

Index • 199

polylines 76, 84, 102, 103, 179polymorphism 10

R

raster datasets 52, 54, 64raster formats 158raster functions 155raster operators 155raster pyramids 156rasters 8, 25, 38, 51, 148, 190,

192multiband 154single-band 154transformation 156

ratio data 152rectification 156reference data 170, 176relational databases

14, 17, 47, 118relationship classes 64, 80, 92

attributed 94cardinality 94notifications 94path labels 94

relationship rules 77relationships 64, 77, 92, 188

cardinality 92composite 94simple 94

renderersbi-unique value 35class breaks 33–34graduated symbol 34proportional symbol 35raster 39simple 32TIN 41unique value 33

resampling 156rings 84, 104routes 179row locks 119row states 120rows 14, 82rubber sheeting 55

S

satellite images 25, 148scanned maps 148segments 103shapefiles 68short transactions 118simple edge features 134simple junction features 134slope 43, 53, 151software interfaces 96solvers 142spatial domains 84spatial indexes 82spatial patterns 26spatial references 80, 84, 173spatial representation 184spectral data 38, 149stretch 40Structured Query Language

12, 14, 28styles 70surfaces 25, 190, 192

comparing raster and TIN 162functional 166

symbols 30fill 31graduated 34line 30marker 30

T

tables 14, 82terrain analysis 151text. See annotationthematic data 38, 149TIN datasets 64topological operators 112topology 16, 77tracing 143triangulated irregular networks

8, 41, 51, 162, 192defined 164

TrueType font 30type inheritance 96

U

unified data model 10Unified Modeling Language

14, 16, 19, 188

V

validation rules 8, 77, 78, 88attribute 88connectivity 88, 133relationship 88

value attribute table 54value field 33versioning 62

conflict classes 123conflict resolution 122–123defined 116fundamentals of 120–121posting 123reconciliation 122

visibility 151Visual Basic 14, 19, 21Visual Basic for Applications

14, 21Visual C++ 14, 21

W

work flows 124work-flow scenarios 116workspaces 66

Z

z values 104, 166ZIP Codes 178

Documents

U.S. Government Restricted/Limited Rights - Esri Supportdownloads2.esri.com/.../ao_/Modeling_our_World.pdf · Modeling Our World The ESRI Guide to Geodatabase Design ISBN 1-879102-62-5