View
4
Download
0
Category
Preview:
Citation preview
Yannis TzitzikasLecture : 12 Date : 22-11-2005
HY351 - Ανάλυση και Σχεδίαση Πληροφοριακών Συστηµάτων
CS351 - Information Systems Analysis and Design
Moving on Design
Design StrategiesLayering and Use of Packages
University of Crete, Fall 2005
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 2
Outline
• The Design Phase (refresher)
• Design Strategies
– Custom Development, Packaged Software, Outsourcing
– Selecting a Design Strategy
• From Analysis to Design
– From Analysis Models to Design Models
– Partitioning the Problem - Layering
– Packages and Package Diagrams
– Packaging and Clustering
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 3
Refresher: The Design Phase
• How the system will operate in terms of
– hardware, software and network infrastructure, the UI, forms and reports, databases,
files, etc
• Steps:
– 1. Design Strategy. Developed by company itself or outsourced to another firm, or
buy an existing software package
– 2. Architecture Design: hardware, software and network infrastructure, the UI, forms
and reports, databases, files, etc
– 3. Database and file specifications: define exactly what data will be stored and where
– 4. Program design: programs that need to be written and what each will do.
• Outcome:
– System Specification that is given to the programming team for implementation
Planning
Analysis
Design
Implementation
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 4
Refresher: Design phase and Development Methodologies
Planning
Analysis
Design
Implementati
on
System
Planning
Analysis
Design
Implementation
System
Design
Implementation
Design
Implementation
Design
Integration
Planning
Analysis
Implementatio
n
Design
Analysis
system v1
Implementatio
n
Design
Analysis
system v2
Implementatio
n
Design
Analysis
system v3
Planning
Implementation
Design
Analysis
System
prototype
Implementation
system
Planning
Implementation
Design
Analysis
design
prototype
Implementation
system
Analysis
Design
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 5
The Design Phase>Design Strategy
• How the system will operate in terms of
– hardware, software and network infrastructure, the UI, forms and reports, databases,
files, etc
• Steps:
– 1. Design Strategy. Developed by company itself or outsourced to another firm, or
buy an existing software package
– 2. Architecture Design: hardware, software and network infrastructure, the UI, forms
and reports, databases, files, etc
– 3. Database and file specifications: define exactly what data will be stored and where
– 4. Program design: programs that need to be written and what each will do.
• Outcome:
– System Specification that is given to the programming team for implementation
Planning
Analysis
Design
Implementation
Design Strategies
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 7
Three Main Design Strategies
• Custom Development
Ανάπτυξη κατά παραγγελία
Πακέτο Λογισµικού
Εξωτερική ανάθεση
Building the system from scratch
Buying a packaged software and customizing it
Hiring an external person/company to build the system
• Packaged Software
• Outsourcing
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 8
Design StrategiesCustom Development
Custom development means building the system from scratch
Advantages
• Complete control (and freedom) over the functionality of the system
• Extensibility of the system (future needs may be taken into account)
• The skills of the staff improve (beneficial for future projects)
Weaknesses
• Takes time, requires hard work (the staff may be already overcommitted)
• Appropriately skilled staff is needed
• High risk
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 9
Design StrategiesPackaged Software
Buying a packaged system and customizing it
(commercial off-the-shelf (COTS) software product)
• many business needs are common to a large number of organizations
• there are thousands of commercially available programs
– that support basic and ubiquitous needs (e.g. payrolls, orders, inventory, sales, etc)
• these programs range from small components to complete ERP (enterprise resource planning) systems
– complete ERP systems have taken years and have cost millions of Euros to develop (SAP, PeopleSoft, Oracle, Baan)
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 10
Design StrategiesPackaged Software (II)
This decision strategy introduces an additional phase, the
“procurement phase” which is a special decision analysis phase comprising the following steps:
1) Identify products that could fit the needs of the project
– sources: special magazines, journals, Web sites, other organizations with similar
needs
2) solicit, evaluate and rank vendor proposals
3) select and recommend the best of them
4) contract with the awarded vendor to obtain the product
Procurement phase: φάση προµήθειας
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 11
Design StrategiesPackaged Software (IΙΙ)
2) solicit, evaluate and rank vendor proposals
Commonly, this step involves preparing an RFQ or an RFP document:
RFQ (Request for Quotations) (αίτηση προσφοράς)
used when you have selected the product but the product is available from various distributions. Objective: negotiate configuration, prices, maintenance contract.
RFP (Request for Proposals) (αίτηση προτάσεων)
used when there are several different vendors and products that are candidate. RFP is a superset of RFQ because the former describes in detail the requirements of the project (it is like the Requirements Specification Document)
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 12
Design StrategiesPackaged Software (IV)
Advantages
• It takes a short time to buy and install it (no extra staff is needed)
• Low risk (packaged software has already been written, tested and proved to be effective)
Weaknesses
• Not full control over the functionality of the system (we cannot customize everything)
– sometimes the business processes of the organization have to adapt to the software!
• Extensibility of the system for the future needs of the organization is not ensured
• The skills of your staff do not improve
• The price of the packages software is sometimes high
• The organization depends on the vendor
– maintenance (plus cost of maintenance)
– Risk if the vendor bankrupts
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 13
Design StrategiesOutsourcing
Hiring an external person/company to build the system
Types of contracts:
• time and arrangements
– payment based on the time and the expenses
– [-] the bill may climb high
• fixed-price contract
– requirements should be very clear
– [-] usually outsourcers are not willing to accept changes in the requirements
• value-added contract
– the outsourcer will have some share of the benefits from the system
– becomes more and more popular
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 14
Design StrategiesOutsourcing (II)
Advantages
• You have the full control over the requirements
• The external party may be more skilled than you
Weaknesses
• Not full control over the internal architecture of the system
• The skills of your staff do not improve
• You may have to share confidential information
• You need a person to manage the outsourcing
• Risk if you haven’t specified well the requirements
– never outsource something you don’t understand
• Risk if the external party fails
Selecting a Design Strategy
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 16
Selecting a Design Strategy
Custom Development
Unique business need
Your staff has technical skills
and you want to enhance
them
Loose and flexible deadline
Packaged System
Very common business need
Technical skills are lacking and it
is not strategically important
to build them
Short timeframe with strict
deadline
You want to avoid risk
Outsourcing
The business need is not crucial
for the organization
Technical skills are lacking and
it is not strategically
important to build them
A “good” outsourcer is available
What about your project?
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 17
Selecting a Design Strategy
Matrix of Alternatives
Description
Operational feasibility
Technical feasibility
Economic feasibility
Cost
Payback period
ROI
Schedule feasibility
Custom
Development
Packaged
Software
Outsourcing
Plus whatever other
criteria are appropriate
We can use a matrix of alternatives (like the one we used in feasibility analysis)
for organizing the pros and cons of each alternative. We could also employ a
weighting scheme or any other decision theoretic method for selecting the best
strategy for the project at hand
Hereafter we assume that we have selected the custom development design strategy
From Analysis Models to Design Models
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 20
From Analysis Models to Design Models
Purpose
• The analysis models we have created so far have mainly focused on FRs
• We should now take into account the NFRs too
• In OO: we have to review and refine the models we have created so far
– by adding system environment details
NonFunctional Requirements (NFR)
• Describe behavioral properties of the system, in terms of characteristics of the form:
– performance, usability, security, legislative, privacy
• Describe how well the system should support the functional requirements
• We could also consider them as “constraints” that restrict the ways that we could
use for implementing the FRs
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 21
From Analysis Models to Design Models
High level
business requirements
described in the system request
Detailed list of more
precise requirements
Functional/Structural/Behavioral
modeling of the system
Requirements
determination
modelingAnalysis
Design
Design Models
Review
Refine,
Account NFRs
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 22
From Analysis Models to Design Models
High level
business requirements
described in the system request
Detailed list of more
precise requirements
Functional/Structural/Behavioral
modeling of the system
Requirements
determination
modelingAnalysis
Design
Design Models
• Class and Method Design
• Data Management Layer Design
• Human Computer Interaction Design
• Physical Architecture Layer Design
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 23
From Analysis to Design Technically, the Design Phase involves:
• Class and Method Design
• Data Management Layer Design
• Human Computer Interaction Design
• Physical Architecture Layer Design
Each of the above topics will be covered in the subsequent lectures
• class specifications (detailed
Class Diagrams)
• Contracts, Method Specifications
• file structure, database
• Use Scenarios, Windows
Navigation Diagrams, Interface
Standards, User Interface
Templates
• Deployment Diagrams,
Hardware/Software
Specifications
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 24
One of the oldest questions in software methods
How to break a large system ?
Reasons: if big then hard to understand, extend, update
Analysis Phase:
In the past: Functional Decomposition (used the days where process and data were separated)
today: Use Cases
Design Phase:
Identify layers and packages (packages are sometimes called subsystems)
Packages are vital tools for large projects
Packages are also good for testing (testing in package basis)
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 25
Identification of Layers
Foundation Layer
Physical Architecture
Human Computer Interaction
Data Management
Problem Domain What we have seen so far
Persistence of objects
User Interface classes
Communication, software architectures
e.g. fundamental data types (usually
Included with the oo development environment)
layers can be represented using the higher level (UML) concept of PACKAGE
An example set of layers:
• So far we have focused on the problem domain layer (else called conceptual perspective)
• Now we have to add system environment information
• One way to avoid overloading the designer is to use layers
Packages and Package Diagrams
Divida et impera (divide and rule)
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 27
Package
A package is a grouping construct that allows us to take any
construct of UML and group its elements into higher-level units.
– It is commonly used to group classes
A package can group
• classes together into higher-level units
• design models, ..
• …….
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 28
Package Diagrams
Dependencies:
• Dependencies are not transitive
• A good practice is to remove dependency cycles (at least reduce them)
• The art of large-scale design: minimize dependencies
P2 P1
P2 depends on P1 if a change to P1 requires changing P1
shows packages and dependencies between them
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 29
A package diagram (grouping classes)
Order Capture
UI
Order Capture
Application
AWT
Mailing List
UI
Orders Customers
Mailing List
Application
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 30
Ways of showing packages on diagrams
java::util::Date
util
package name
Date
util util
Date
package name + contents
java::util
Date
fully qualified
package name
util
Date
java
nested packages
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 31
Packages, Visibility and Importing
• We can define the visibility of the elements owned by a package as we defined the visibility of the attributes and operations of a class.
• +: public
• -: private: seen only the elements of the same package
• #: protected: seen only by the elements of the same
package and its children
• To refer to the class of another package you need to use the qualified name of the class (package name::class name)
• import allows using direct names (not qualified ones)
• importing adds the public elements from the target package to the public namespace of the importing package. Notation in UML: <<import>>
“Import” vs “Access”: “Access” is like “import” but you cannot reexport. So it is not transitive (import is transitive)
util
+Date
+Tree
- XYZ
P2 P1<<import>>
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 32
More Examples
presentation<<subsystem>>
control<<subsystem>>
entity<<subsystem>>
mediator<<subsystem>>
foundation<<subsystem>>
Cinterface
Minterface
Einterface
Finterface
java
applet
awt
lang
Layers as Packages
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 34
Modeling Layers using Packages
HCI
Data
Management
Problem
Domain
Foundation
Physical
ArchitectureFoundation Layer
Physical Architecture
Human Computer Interaction
Data Management
Problem Domain
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 35
Packages for Modeling the various Architectural Views
We could view packages to model the views of an architecture
Design ViewImplementation
View
Use Case View
Deployment
ViewInteraction View
Data Storage
Data Access Logic
Application Logic
Presentation Logic
Layer N
Layer N-1
Layer 3
Layer 2
Layer 1
Packaging Classes
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 37
Packaging classes: Motivation
• Classes is the basic means for structuring an object-oriented system.
• But is this structuring enough if we have hundreds of classes ?
Tip: Use packages whenever a class diagram does not fit in an A4 paper
NO TE: The relationsh ips in blue a re generic re la tions hips that are de tailed by spe cific relationsh ips .
0-1
0-N
W orkers Com pe nsa tionClaim-PolicyHolder
1 -1
0 -NWorkers Compe ns ationCla im -Notifie r
1-1
0-N
W ate rVeh ic le-Sea tingCa pa city
1-1
0-NWate rVeh icle-LoadCapa city
1-1
0 -NWa terVe hicle -Eng ine
1 -1
0-N
Victim-Pro fes sionSen io rity
1 -1
0-1
Vic tim-MutualHe alth Ins u rer
1 -1
0-N
Vic tim -JobCa tego rySe nio rity
1 -1
0 -N
Vic tim-Com pa nySe nio rity
1 -1
0-N
Vic tim -BranchSen iority
1-1
0 -N
Victim-Bonus Rec ep tionDe cla rati
0 -N
1-1
Victim-A IKRe cep tionD ec la ration
1 -10 -1
Vic tim -Ac tua lIn s ure r
1 -1
0-N
Victim-A ctiv ityExe cu tionAddres
1 -1
0 -1VERIDA SSRe gis tration-Re que ste r
0 -1
0-N
Veh icle-Tra nsporte dGoods
0-1
0 -N
Veh ic le-Sec u ritySys tem
1-1 0-1Veh icle-Re gis trationPlate Ho lde
1-1
0 -1
Veh icle-Moto rTPL
1-1
0 -1
Ve hic le -MotorTheft
1 -1 0 -1Veh icle-Moto rPartialComprehens ive
1-1
0-1
Veh icle -Moto rFu llComprehens ive
1-1
0-1
Veh ic le-Moto rFire
1-1 0-1Ve hic le -Lea singCompany
1-1
0-N
Tota lLos s-Re ques tForBid
0 -1
0-N
Subguara nte e-Cover
1 -1
0-N
Spec ia lCond ition -Cla use
0-1
0-NSettlem en t-Pa rty
0 -1
0 -N
Se ttleme n t-Gua ran tee
0-1
0-N
Se ttleme n t-Cover
0 -1
1-1 RiskObjec t-Pre viousInsu re r
1-1
0-N
Reques tForBid -Bid /Offe r
0-1
0-NRela tedParty -Rate
1 -10 -1 Rela tedParty -OfficialA ddre ss
1-1
0-N
Re la tedParty -Office Tele phone
1-1
0 -N
Re la tedParty -MobileTe lephone
1 -1
0 -N
Re la tedParty -Home Telephone
1 -1
0 -N
Re la tedParty -Fax
1-1
0-N Re late dPa rty -Exte rna lRole
1 -1
0-N
Re la tedParty -eMa il
1-1
0 -N RegLegPers on -PartyRo le
1-1
0-1
Re gLe gPe rs-Soc ialSec urityBure a1 -1
0 -1
Re gLe gPe rs-ParityCommittee
1-1
0 -1
Re gLegPers -Ho lida yPayBurea u
1 -1
0-1
RegLegPers -FamilyA llowa nce Bure au
1-1
0-1Property -RiskAddres s
0 -1
1 -1
Pro fes sion-A ctiv ityExe ctu tionAddres s
1-1 0 -NPrivate Veh icle/Truck -Driver
1-1
0 -N
Pre Not-Renewa lPre miumPeriod
1-1
0-N
PreN ot-Pro rationPre mium Pe riod
0-N
0-1
0-N
Pre miumNotific ation-PIRe la tedA
1-1
0-1
Prem iumLine-Po licyHolder
1-1
0-1
Pre mium-Tax
1-1
0-1
Pre mium -Subsc ription Index
1-1
0-1
Pre mium-SplitPa ymen tLoad ing
1-1
0-1
Prem ium-Prev ious Inde x
1-1
0 -1Prem ium-O the rIns u rance Reductio
1-1
0 -1
Prem ium-FidelityBonus
1 -1
0 -1
Premium-Comme rcialDisc oun t
1 -1
0 -1
Prem ium-A ctualIndex
1-1
0-N
PolicyHolder-RSRReg istra tion
0-N
0-1
0-N
PolicyHolder-Prev ious Po licy
0-N
0-1
Po licyHolde r-Pre v iousCla im
0-1
0-N
PolicyHolder-Prev InsRe fus al
0 -N
1-1
Po lic yHolde r-Pre ve n tionMea su re
0-1
0 -N
PolicyHolde r-Offe nc ePend ing
0 -1
0-N
PolicyHolde r-NCBLeve l
0 -1
0-N
PolicyHolder-Conviction
0-1
0-1
Po lic yH older-BillingAddres s
0 -1
0 -N
Po lic yH older-A cc ide nts Sinc eAdjustme n t
1 -1
0 -N
Phys ica lPers on-Pa rtyRo le
1-1
0 -N
Phys icalIn jury -Tre atingMe dic al
1 -1
0-N
Physica lInju ry-H os p ita l
1 -1
0-N
Phys ica lIn jury -FirstA idSupp lie
1-1
0-N
Physica lInju ry -Disab ility
1-1
0-N
Physic alIn ju ry -Certify ingMed ic
1 -1
0 -1
Party -Pre v iousA ddre ss
1-1
0-1
Party -Officia lA ddre ss
1-1
0 -N
Pa rty -Offic eTe lephone
1 -1
0-N
Party -Mob ileTe le phone
1-1
0-1
Party -Ma ilPrim aryAddres s
1-1
0-N
Party -HomeTe lephone
1-1
0-N
Pa rty-Fa x
1-1
0-N
Pa rty -eMail
1-10 -N Oppone n tThirdPa rty -Circ umsta nc eCo lumn
1-1
0-N NonRegLegPe rson -PartyRole
1 -1
0 -1
Motoris edRoadVeh icle-Ve h ic lePas se nge r
1-1
0 -N
Moto ris edRoadVeh icle -Tec hn ica rData
1-1
0-1
Moto ris edRoadVe hic le -Se atingCapac ity
1-1
0-1
Moto rise dRoa dVe h ic le-LoadCapac ity
1-1
0-N
Moto ris edRoadVeh icle -Hau lageTra ile r
1-1
0-1
Motoris edRoadVeh icle-Eng ine
1-1
0-N
Moto ris edRoadVeh icle -Cara va n
1-1
0-N
Moto rIns u rance Contra ct-VERIDA S
1-1 Mate ria lO bje ct-Ow ne r
1-1 Ma terialOb jec t-Fina nc eCompany
1-1
0-1Ma terialDa ma ge -Repaire r
1-1
0-1
Ma teria lD ama ge-Da mage dO bje ct
1-1
0-NMa terialDa ma ge -As se ss or
1-1
0-N
Life Ins Con tra ct-Re ferenc ePrem ium
1-1
0-N
LifeInsCon trac t-Life Pro fitSha ring
1-1
0-N
Life Ins Contra ct-Las tPremium
1-1
0-N
LifeInsCon trac t-Firs tPre mium
1-1
0-N
Life InsCon trac t-Dec ea se Pro fitSha ring
1 -1
0-1
Le galPers on -He adOffice Addres s
1-1
0-1 Le galPers on -Branc hO ffic eA ddre s
1 -1
1 -1
KfkO pe ration-Se nd ingCompany
1-1
1-1
KfkO pe ration -Re ce ivingCom pa ny
1-1
0 -N
Insu re dPers on-Prima ryO ccupation0-1
0 -1
Insu redPers on -Perm anen tDisa b il
1 -1
0 -1
Insu redPers on -Ge ne ralLife
1 -1
0-N
Ins u redPerson -Add itionalOc cupation 1 -1
0 -N
Insu re dPers on-A dditiona lActiv itie s
1-1
0 -N
Insu red -Circums tanc eCo lumn
1-1
0 -N
Insu ra nc eProvider-FileA dmin ist
1 -1
Inc ide n t-Witne ss
1 -1
Inciden t-Oppone ntTh irdPa rty
1 -1
Inc ide n t-Inte rven ingAutho rity
1 -1 Inc ide nt-Even tA ddre ss
1-1
0 -N
H un ting -Hun tingLice nc e
1 -10 -N Gua ran te e-Subguara nte e
1-1
0 -N
Guara nte e-Spec ia lCond ition
1 -1
0 -1
Guara nte e-Proportiona lPre mium
1-1
0-1
Gua ran tee -OldPre mium
1-1
0-1
Gua ran tee -Ne tPre mium NonIndexed
1-1
0-1
Gua ran tee -NetPre miumIndexed
1-1
0 -N
Guara nte e-In sVa lBas edDeductib le
1-1
0 -1
Guara n tee -Gros sPrem ium
1-1
0-N
Guara nte e-Ge ne ralCond ition
1-1
0-N
Guara n tee -Ex tension
1 -1
0 -1
Gua ran tee -Dam AmBa se dD eductib le
1 -1
0 -N
Guara nte e-D ailyA llowa nc e
0 -1
0-N
Gua ran tee -Cove r
1-1
0 -N
Gua ran tee -Cost/Charge
1-1
0 -N
Guara n tee -Com mission
1 -1
0 -N
Gua ran te e-Ca pita l
1 -1
0-1
Gua ran tee -Bas ePrem ium
1-1
0-1
Guara n tee -Annua lPre mium NonInde
1 -1
0-1
Gua ran tee -AnnualPremiumInde xe d
1-1
0 -N
Guara nte e-A nnua lIndemnifica tion
1-1
0-NGe ne ralLife-Be ne fic iary
1-1
0 -N
Gene ralCond ition -Clause
0-1
1 -N
Flee t-Veh icle
1 -1
0-N
Fac tua lGroup-Pa rtyRole
1 -1
1 -2
EAN Form-Circums tanc eCo lumn
0-10-N Driver-Prev ious Claim
0-1
0 -N
Drive r-NCBLevel
1 -1
0 -N
Drive r-DrivingLic ence Withd ra wa
1-1
0 -1
D river-D riv ingLice nc e
0-1
0-N
Driver-A cc ide nts Sinc eAdjustme n t
1-1
0-N
Docume nt-D oc umen tExcha nge Deta ils
1 -10-N De cla ration -Re spons e
0-1
0-NDa ma ge dObje ct-Va lua tion
0-1
1-1
Dama ge-Guara n tee
1-1
0-N
CurA cc -Pre miumRetu rnLine
1 -1
0 -N
CurAc c-Prem iumRe im burs eme ntLine
1-1
0-N
CurAc c-Pre miumInvo ic eLine
1-1
0-N
CurAc c-Prem iumCa nc ellationLine
1-1
0 -N
CurAc c-Misc ellTrans Line
1-1
0-N
CurA cc-FinTransLine
1-1
0 -N
CurA cc -Com miss ionLine
1 -1
0 -N
CurA cc -ClaimLine
1-10 -N CurAc c-Borde rea uTota lLine
1 -1
0 -1
CoverLimit-Subsc ription Inde x
1-1
0-1
Cove rLimit-Prev ious Inde x
1-1
0 -1
Cove rLimit-Ac tualInde x
1-1
0-N
Coup le-Pa rtyRo le
1-1
0-1
Con trac t-Transfe r
1 -1
0-1
Contra ct-Suspe ns ion
1 -1
0 -N
Con trac t-Spe cialCondition
0-1
0-N
Contra ct-Re la tedPo lic y
1 -1
0-1
Con trac t-ProportionalPremium
0-1
1-N
Con trac t-Pre miumNotific ation
0 -1
1-N
Con trac t-Po licyHolde r
1-1
0-1
Con trac t-PolicyChange
1 -1
0 -1
Con trac t-Ne tPre miumNonIndexed
1-1
0 -1
Contra ct-N etPrem iumIndexe d
1-N
0-N
Con trac t-Inte rme dia ry
1-N
0-N
Con trac t-Ins u rer
1 -N Con trac t-Ins u ra nce Obje ct
1 -1
0 -1
Contra ct-Gros sPremium
1-1
0-N
Con trac t-Genera lCond ition
1-1
0-N
Con trac t-Cos t/Charge
0-N
1-1
Con trac t-CoPo lic yH older
1-1
0 -N
Contra ct-Con trac tDocume nt
1 -1
0 -N
Con tract-Comm is sion
1-1
0 -N
Con trac t-Co insu rer
0-1
0 -N
Con trac t-Cla im
1-1
0-1
Contra ct-Ca nce lla tion
1-1
0-1
Con trac t-Annu lm en t
1 -10 -1 Con trac t-Annua lPre mium NonIndex
1-1
0 -1
Contra ct-A nnualPrem iumIndexe d
1-1
0 -1
Con tra ct-A mendmen t
0-1
0 -N
Con ten ts -Se cu ritySyste m
0-1
0 -1
Conte n ts-Ge ne ralFire
0-1
0-1
Conte n ts-Burg la ry/The ft
1-1
0 -1
ClaimLine -Po licyHolde r
0 -1
0-N
Claima nt-Se ttle men t
0-N
Cla iman t-Pro fess ion
0-10 -N Claima n t-Dam age
0-N
0-1
Claim-Ma na geme n tAct
1-1
0-N
Cla im-KfkOperation
1-1
0-N
Cla im-In terme dia ry
1 -1
0-1
1 -N
Cla im-Claima nt
1 -1
1-6
Circums tanc eCo lumn-Circums tance
0-1 0-NBuild ing-Ta ke OverPo lic y
0 -1
0-1
Bu ild ing -Storm HailAndHea vySnow
0-1
0-N
Build ing -Sec u ritySys tem
0-1
1 -1
Bu ild ing -PolicyHolder
0 -1
0-1 Bu ild ing -Gene ralFire
1-1
1 -N
Borde rea u-Bordere auLine
1-1
0 -N
Bonus Rec ep tionDe cla ration -Bonus
0 -1 A ss es sme n t-Se ttle me nt
1-1
As se ss men t-Da ma ge dObje ct
1-1
0-N
A irVe hic le -Sea tingCapac ity
1-1
0-N
A irVe h ic le-Loa dCapac ity
1-1 0-NAirVeh icle-Eng ine
1 -1
0-N
AIK Rec ep tionDe clara tion -AIK
0-1
0 -1
Ac countingDocume nt-In term ed iar
0-1
0-1Ac countingDocumen t-In su re r
P
Worke rsCompens ationCon trac t
Workers Compensa tionClaim
WorkersCompe ns ation
Witnes s
Wate rVeh icle
Victim
VERID ASSReg istration
Ve h ic leRe gis tra tionPlate Ho lder
Ve h ic leRe gis tra tionCa rd
Ve hic le Pass engers
Ve hic le
Va lua tion
Tre atingMed icalDoc tor
Trave lIn su ra nc eCon trac t
Tra nspo rtIn s ura nce Con trac t
Trans po rtedGoods
Tra nsfe r
To talTe mpora ryD is ab ility
Tota lLos s
Th irdCom mission
Tem poraryDis ab ility
Telephone
Te chn ic arD ata
Ta x
TariffPre mium
Ta ke OverPo lic y
Suspens ion
Subs crip tion Inde x
Subgua ran tee
Sto rmHa ilA ndHe avySnow
Sta rtInde x
SplitPre miumPaymen tLoad ing
Spe cialCondition
Spac eVe h ic le
Socia lSe cu rityBurea u
Settle men t
Se n io rity
Send ingCompany
Sec uritySys temCe rtific ate
Se cu ritySys tem
Se condRes idenc yA ddre ss
Se condComm is sion
Se atingCapac ity
RSRReg istra tion
RoadVe hic le
Ris kObjec t
Ris kA ddre ss
Re sponse
Res ide n tia lCarava n
Re que stFo rBid
Reques ter
Repres en ta tive
Re p la ce dPo lic y
Repa ire r
Renewa l
Rela tedPo licy
Re late dPa rty
Reg iste redLegalPers on
Refere nce Pre mium
Rece ivingCom pa ny
RealEsta te
Rate
RailVeh ic le
Pro ration
ProportionalPremium
Prope rty
Pro fitSha ring
Profe ss ionSen iority
Priva teMoto ris edRoadVe hicle
Private Build ing
PrimaryOc cupation
Pre v iousPo lic y
Pre viousInsu re r
Pre vious Insu ranc eRe fus al
Pre viousIndex
Prev ious Claim
Pre viousA ddres s
Preven tionMe as u re
Pre miumRetu rnLine
Pre miumRe imburs em en tLine
Prem iumPeriod
Pre mium Notific ation
Pre mium Loa ding
PremiumLine
PremiumInvo ice Rela tedAc tion
Pre miumInvo ic eLine
Pre mium Invo ic e
PremiumDis coun tOrLoad ing
Pre miumDisc oun t
PremiumCanc ellationLine
PremiumA ndCommiss ionRe ve rsib le PremiumLine
Prem ium
Polic yH older
Po lic y
Physica lPe rson
Phys ica lIn ju ry
Pe rsonalLia bilityIns ura nceCon trac t
Pe rma ne ntD is ab ilityCove r
Perma nen tDisa bility
PartyRole
Pa rty
PartialTem pora ryD is ab ility
Pa rityCommittee
Owner
OtherPriva teBu ild ingO the rPolicy
O the rInsu rance Reduction
Oppone n tTh irdPa rty
OldPre mium
OfficialA ddre ss
Office Telephone
O ffence Pending
Notifie r
NonRe g is tere dLe ga lPe rson
NonMoto rise dRoa dVe h ic le
N etPremiumNoninde xed
Ne tPre miumInde xed
N etPrem ium
NCBYe arMinus 3
NCBYea rMinus 2
NCBRefe renc e
N CBPre v iousNCBLevel
N CBAtUnderwriting
NCBA tPrev iousIns ure r
NCBA ctual
MutualHe alth Ins ure r
Moto rThirdPartyLia bility
Moto rTheft
Moto rPa rtialComprehensive
Motoris edRoadVeh icle
Moto rInsu ra nc eCon trac t
Moto rFullComprehensive
MotorFire
MotorCyc le Certifica te
Moto rComprehens ive
Mortgage Loa nFile
MoralDam age
MobileTe le phone
Mis ce lla neousTrans ac tionLine
Me d ic alDoc tor
Ma terialDa mage
Manage men tAc t
Ma ilSec onda ryA ddre ss
Ma ilPrimaryA ddres s
Ma ch ine ry /App lia nc es
Loa dCa pa city
Life Pro fitSha ring
LifeInsu ranc eCon trac t
Le ga lPe rson
Le ga lExpe nse sInsu ranc eCon trac t
Lea singCompany
La stPre mium
La stCommis sion
Land
K fkRe que stForRe fund
KfkOpera tion
KfkDeb itOrCred itStatem en t
JobCate go rySe n io rity
Inte rven ingAutho rity
In terme d ia ry
Ins u rer
In su re dVa lue Bas edDe duc tib le
Ins u redPers on
Insu red
Insu ra nc eProvider
Indiv idua lIn s ura nceCon trac t
Index
Inde mnifica tion
Ide n tityCard
H un tingLice nc e
Hun ting
H os p ita l
Home Telephone
H olidayPayBure au
Hea dOffic eA ddre ss
H au lageTra ile r
Gua ran tee
GroupIns u rance Con tra ct
Group
Gros sPrem ium
Gre enCa rd
Gene ralLife
Gene ralFire
Genera lCond itionsWord ing
Genera lCond ition
Genera lCiv ilLiab ility
Ga rage
Fu tu reA ddre ss
Flee t
Firs tPre mium
Firs tCom miss ion
Firs tAidSupp lie r
FireInsu ra nc eCon trac t
Financ ia lTra nsa ctionLine
Financia lD ama ge
Financ eCompany
File Adminis trato r
Fide lityBonus
Fee
Fax
Fa milyA llowa nc eBurea u
Fac tualGroup
Exte rna lRole
Ex tension
Expe nse
Eve ntA ddre ss
Europe anA cc ide ntNo tifica tionForm
Eng ine
EndInde x
Employe eBe nefitsCon trac t
e Mail
D welling
D riv ingLice nc eW ithd raw al
DrivingLic ence
D river
Docume n tExchangeDe tails
Doc ume n t
Disa bility
D ire ctPayme n t
D educ tible
D ec la ration
De ce as ePro fitSharing
Da taEn tryO pe rato r
D ama gedObject
Dam ageA mountBa se dD eductib le
Dam age
DailyA llowa nce
Curre n tAc countLine
Curren tAc coun t
Cred itor
Cove rNo te
CoverLimit
Cover
Couple
Cos t/Cha rge
CoPo lic yH older
Conv iction
Con trac tDocume nt
Con trac tChange
Con trac t
Con ten ts
Cond ition
Com pa nySe nio rity
Communic ationCha nne l
Com monLaw
Commis sionLine
Commiss ion
Com merc ialMoto rise dRoa dVe h ic le
Comme rcia lLoss
Comme rcia lIn s ura nc eCon trac t
Com merc ialDisc oun t
Comm erc ia lBuilding
Co lle ction
Co insu rer
Clause
ClaimLine
Claima nt
Claim
Civ ilLiab ilityHorse sA ndCa rriages
Circ um sta nc eCo lum n
Circums tance
Ce rtify ingMe dic alDocto r
Carava n
Cap ital
Ca nc ellation
Ca lc u la tion
Burg la ry /The ft
Bu ild ing
BranchSen io rity
BranchOffice Addres s
Borde rea uTo ta lLine
Borde rea uLine
Borde rea u
Bonus Rec ep tionDe clara tion
Bonus
BillingA ddre ss
Bid /Offe r
Bic yc le
Be neficia ry
Bas ePremium
A ss es so r
Annu lme nt
A nnualPrem iumN onInde xe d
Annua lPre miumInde xed
Annua lPre mium
A nnualIndemnification
Am endmen t
A irVe h ic le
Advan tage InKindRec ep tionDe clara tion
A dva nta geInKind
Addres s
AddO nIns ure r
A dd itiona lO cc upation
Add itiona lAc tiv ity
A ctua lIn s ure rOf
A ctualIndex
Ac tivityExec utionAddres s
A ction
Acqu isitionCommiss ion
Acc oun tingD oc umen tLine
Ac coun tingDocume nt
A cc iden ts Since Adjustme nt
Overv iew /Conc ep tua l
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 38
Packaging classes: Properties
Some properties :
• a class can be member of a single package
• a class should have a unique name within the package
• a package can contain both classes and packages
So we can define an hierarchic structure
Notes:
• UML packages correspond to packages in Java)
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 39
Packaging Classes
The packaging of classes can be done according to several different criteria, e.g.:
• by authorship (who wrote what)
• by functional relationship (e.g. Storage Manager, UI, MVC, …)
• by architectural tier, e.g. we could distinguish the following types of classes:
Data Storage
Data Access Logic
Application Logic
Presentation Logic
Other distinction:
•persistent database classes
•entity classes: those that will reside in the
main memory
•boundary classes (interfaces)
•control classes: specify business logic function
Package Diagram of Appointment System
Data Storage
Data Access Logic
Application Logic
Presentation Logic
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 41
Which classes to put in which packages ?
Two general principles:
• Common Closure Principle
– the classes in a package should need changing for similar reasons
• Common Reuse Principle
– the classes in a package should all be reused together (semantically close)
These principles drive us to the notion of cohesion
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 42
Packaging classes: Hints and Tips
A well structured package
• is cohesive (providing a crispy boundary around a set of related elements)
• is loosely coupled (exporting only those elements other packages really need to see and importing only those elements necessary and sufficient for the elements in the package)
• is not deeply nested (human understanding is limited)
Cohesion =συνεκτικότητα/συνοχή, Cohesive = συνεκτικό
coupling = σύζευξη, loosely coupled = χαλαρά ζευγµένα
Packaging relates to clustering
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 43
Clustering
• Clustering is the process of grouping similar objects into naturally associated subclasses.
• This process results in a set of “clusters” which somehow describe the underlying objects at a more abstract or approximate level.
• The process of clustering is typically based on a “similarity measure” which allows the objects to be classified into separate natural groupings.
• A cluster is then simply a collection of objects that are grouped togetherbecause they collectively have a strong internal similarity based on such a measure.
• A similarity measure (or dissimilarity measure) quantifies the conceptual distance between two objects, that is, how alike or disalike a pair of objects are.
– Determining exactly what type of similarity measure to use is typically a domain
dependent problem.
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 44
Clustering
• Clustering is used in areas such as:
– medicine, anthropology, economics, soil analysis, data mining
– software engineering (reverse engineering, program comprehension, software
maintenance)
– information retrieval
• In general, any field of endeavor that necessitates the analysis and comprehension of large amounts of data may use clustering.
A clustering of a set N is a partition of N, i.e. a set C1,…, Ck of subsets of N, such that:
C1 ∪… ∪ Ck = N and Ci ∩ Cj =∅, for all i≠j.
Packaging relates to graph clustering
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 45
Graph Clustering
• Graph clustering deals with the problem of clustering a graph
– grouping similar nodes of a graph into a set of subgraphs
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 46
Quality criteria for graph clustering methods
Graph clustering methods should produce clusters with
high cohesion and low coupling
• high cohesion:
– there should be many internal edges
• low “cut size”:
– The cut size (else called external cost) of a clustering measures how many edges are
external to all sub-graphs, that is, how many edges cross cluster boundaries.
• Uniformity of cluster size is also often desirable.
– A uniform graph clustering is where |Ci| is close to |Cj| for all i,j in {1..k}
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 47
Example
1
2
3
4
5
6
7
8
9
1
2
3
4
5
6
7
8
9
� �
�
Cut size =2
1
2
3
4
5 6 7
89
Cut size =4 ��
�
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 48
Quality Measures for Graph Clustering
• There are several. One well known is the CC measure (Coupling-Cohesion measure)
||
||||
E
EECC
exin−
=
• Ein: the “internal” edges: those that connect nodes of the same cluster
• Eex: the “external” edges: those that cross cluster boundaries
• maximum value of CC: 1
– when all edges are internal
• minimum value of CC: -1
– when all edges are external
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 49
Example
1
2
3
4
5
6
7
8
9
� �
�
Cut size =2
1
2
3
4
5 6 7
89
Cut size =4 ��
� 2.010
46=
−=CC
6.010
28=
−=CC
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 50
Hierarchical Graph Clustering
• The clusters of the graph can be clustered themselves to form a higher level clustering, and so on.
• A hierarchical clustering is a collection of clusters where any two clusters are either disjoint or nested.
�
1
2
3
4
5
6
7
8
9
�
�
�
�
1
2
3
4
5
6
7
8
9
� �
�
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 51
Hierarchical Clustered Graph
A Hierarchical Clustered Graph (HCG) is a pair (G,T) where
G is the underlying graph, and
T is a rooted tree such that the leaves of T are the nodes of G.
(the tree T represents an inclusion relationship: the leaves of T are nodes of G, the internal
nodes of T represent a set of graph nodes, i.e. a cluster)
1
2
3
4
5
6
7
8
9
� �
�
�
�
1
23
4
5 6 7
8
9
A B C
D
EHCG
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 52
Implied Edges
Implied edges: edges between the internal nodes.
Two clusters are connected iff the nodes that they contain are related.
Multiple implied edges (between the same pair of clusters) can be ignored or summed up to form weighted implied edges. Thresholding can applied in order to filter out some implied edges
1
23
4
5 6 7
8
9
A B C
D
E
A Hierarchical Compound Graph is a triad (G,T, I)
where (G,T) is a hierarchical clustered graph (HCG), and I the set of implied edges set.
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 53
Packages and Graph Clustering
Which is the graph G=(N,E) in our case?
• N: classes
• E: various dependency relationships
– structural, e.g. associations
– behavioral, e.g. the edges of communication diagrams
– “semantical” (similarity of use/purpose: hard to infer automatically]
Which is the hierarchical compound graph (G,T, I) in our case?
• G (as before)
• T
– leaves of T: nodes of N (i.e. classes)
– internal nodes above leaves: packages
– internal nodes of higher level: packages of packages
• I : the implied edges in our case are the dependency relationships between
the packages
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 54
Packaging and Clustering: Homework
• Assume we have an implemented system whose classes are not packaged.
• How could you partition them to packages?
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 55
Algorithms for Graph Clustering
• Node similarity-based algorithms
• Kernighan-Li and variant algorithms
• Graph growing algorithms
• Spectral σ-section algorithms
• Multilevel partitioning algorithms
• ...
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 56
Summary
• Alternative strategies to design and create the new system: Custom building, packaged
software, outsourcing
• The object-oriented approach to analysis and design is based on use-case modeling, is
architecture centric, and supports functional, static and dynamic views of the system.
• When evolving analysis into design models, it is important to review the analysis models
then add system environment information.
• Packages and package diagrams help provide structure and less complex views of the
new system.
– Set the context
– Cluster classes together based on shared relationships
– Model clustered classes as a package
– Identify dependency relationships among packages
– Place dependency relationships between packages
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 57
Reading and References
• UML Distilled: A Brief Guide to the Standard Object Modeling Language (3rd Edition)
by Martin Fowler, Addison Wesley, 2004.
• The Unified Modeling Language User Guide (2nd edition) by G. Booch, J. Rumbaugh, I.
Jacobson, Addison Wesley, 2004
• Systems Analysis and Design with UML Version 2.0 (2nd edition) by A. Dennis, B. Haley Wixom, D.
Tegarden, Wiley, 2005. CHAPTER 9
• System Analysis and Design Methods (6th edition) by Jeffrey L. Whitten, Lonnie D.
Bentley and Kevin Dittman, McGraw-Hill, 2004, Chapter 12
How to break a large system ?if big then hard to understand, extend, update
How to break an existing system in order to understand it?A relevant method is presented next.
How to Tame a Very Large ER Diagram(using Link Analysis and Force-Directed Placement Algorithms)
Yannis Tzitzikas1 and Jean-Luc Hainaut2
1University of Crete & FORTH-ICS
2University of Namur (F.U.N.D.P.), Belgium
24th Intern. Conf. on Conceptual Modeling, ER’2005, Klagenfurt, Austria, Oct. 2005U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 60
How to Tame a Very Large ER Diagram
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 61
Very small excerpt of a
schema (DICO1) with
456 entity and
812 relationship types
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 62
Conceptual Diagrams
• They are useful and widely used in
– Requirements Engineering
– Reverse Engineering
– Knowledge Representation
• Examples
– ER diagrams
– UML class diagrams
– SWeb ontologies
– ...
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 63
But why we use them ?
They aid communication• customers vs analysts
• analysts vs domain experts
• analysts vs designers
• designers vs designers
• designers vs developers
• designers vs end users
• ...
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 64
But why we use them ?
They aid communication• customers vs analysts
• analysts vs domain experts
• analysts vs designers
• designers vs designers
• designers vs developers
• designers vs end users
• ...
create table Component (
C_P_ID_Par char(10) not null,
ID_Par char(10) not null,
Quantity char(1) not null,
constraint ID_Component primary key (C_P_ID_Par, ID_Par));
create table Deparment (
ID_Dep char(10) not null,
DepId char(1) not null,
DepName char(1) not null,
Address char(1) not null,
constraint ID primary key (ID_Dep));
create table Dependent (
FirstName char(1) not null,
LastName char(1) not null,
YearOfBirth char(1) not null,
Supporter char(10) not null);
create table Employee (
ID_Emp char(10) not null,
EmpId char(1) not null,
FirstName char(1) not null,
LastName char(1) not null,
MiddleName char(1) not null,
YearOfBirth char(1) not null,
Salary char(1) not null,
ID_Dep char(10),
constraint ID primary key (ID_Emp));
create table Part (
ID_Par char(10) not null,
PartNo char(1) not null,
PartDescription char(1) not null,
QuantityOnHand char(1) not null,
constraint ID primary key (ID_Par));
-- Constraints Section
-- ___________________
alter table Component add constraint FKconsistsOfforeign key (ID_Par)references Part;
alter table Component add constraint FKCom_Parforeign key (C_P_ID_Par)references Part;
alter table Dependent add constraint FKEmp_Depforeign key (Supporter)
references Employee;alter table Employee add constraint FKDept_Emp
foreign key (ID_Dep)references Deparment;
alter table Project add constraint FKProj_Managerforeign key (ID_Emp)references Employee;
alter table Proj_Work add constraint FKPro_Proforeign key (ID_Pro)
references Project;alter table Proj_Work add constraint FKPro_Emp
foreign key (ID_Emp)references Employee;
alter table Supp_Part add constraint FKSup_Sup_1
foreign key (ID_Sup)references Supplier;
alter table Supp_Part add constraint FKSup_Par_1foreign key (ID_Par)references Part;
alter table Supp_Part_Proj add constraint FKSup_Supforeign key (ID_Sup)references Supplier;
alter table Supp_Part_Proj add constraint FKSup_Proforeign key (ID_Pro)
references Project;alter table Supp_Part_Proj add constraint FKSup_Par
foreign key (ID_Par)references Part;
create table Project (ID_Pro char(10) not null,ProjId char(1) not null,Title char(1) not null,ID_Emp char(10),
constraint ID primary key (ID_Pro));create table Proj_Work (
ID_Emp char(10) not null,ID_Pro char(10) not null,timePercentage char(1) not null,
constraint ID_Proj_Work primary key (ID_Pro, ID_Emp));
create table Supplier (
ID_Sup char(10) not null,SupId char(1) not null,Name char(1) not null,Status char(1) not null,Address char(1) not null,
constraint ID primary key (ID_Sup));create table Supp_Part (
ID_Par char(10) not null,ID_Sup char(10) not null,
constraint ID_Supp_Part primary key (ID_Par, ID_Sup));
create table Supp_Part_Proj (
ID_Par char(10) not null,ID_Pro char(10) not null,ID_Sup char(10) not null,Quantity char(1) not null,constraint ID_Supp_Part_Proj primary key
(ID_Par, ID_Sup, ID_Pro));
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 65
But why we use them ?
They aid communication• customers vs analysts
• analysts vs domain experts
• analysts vs designers
• designers vs designers
• designers vs developers
• designers vs end users
• ...
A picture is worth a thousands of words
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 66
However, their usefulness degrades rapidly as they grow in size
Very small excerpt of a
schema (DICO1) with
456 entity and
812 relationship types
A conceptual diagram with thousands of elements
does not worth a lot.
(unless it has a
a regular form)
Big schemas become more and
more frequent
– SAP database consists of
about 30.000 tables
– Semantic Web applications
– Reverse engineering
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 67
The objective:
E.g. generation of more abstract or focused views
ConfiguredComputer
configuredPrice : float
StandardComputer
standardPrice : float
ConfigurationItem
itemType : StringitemDescr : String
Customer
customerName : StringcustomerAddress : StringphoneNumber : StringemailAddress : String
Payment
paymentMethod : StringdateReceived : DateamountReceived : float
Invoice
invoiceNumber : StringinvoiceDate : DateinvoiceTotal : float
Computer
computerName : String
1..n 1..n
Order
orderNumber : StringorderDate : DateshipAddress : StringorderTotal : floatorderStatus : StringsalespersonName : String
0..n
1..1
0..n
1..1
1..1
1..1
1..1
1..1
1..1
0..1
1..1
0..1
1..n 0..n1..n 0..n
Device techniques
to aid the understanding and the visualization
of large conceptual diagrams
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 68
Existing Solutions for Managing Large Diagrams (1/3)[1] ER Clustering
– The classical methods require human input
• [Felldman&Miller’86,Teory et al.’89, Gandhi et al.94, Campbell et al.’96]
– Automatic methods not tested in large ER diagrams
• [Akoka&Comyn-Wattiau’96, Raugh&Stickel’92]
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 69
Existing Solutions for Managing Large Diagrams (2/3)[2] Graph Drawing and Abstraction
– Automatically generated layouts are not satisfying
• Most (if not all) of the layouts are still created manually
– Hierarchical decomposition techniques that are used for visualizing big plain graphs have not been applied or tested on conceptual diagrams
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 70
Existing Solutions for Managing Large Diagrams (3/3)[3] Visualization methods
– Not very helpful for our problem
• Hyperbolic trees: appropriate only for trees
• FishEye View: not really helpful and computationally hard
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 71
The Idea
Try to rank the elements of the diagram
according to their “importance”
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 72
Why ranking can be useful?
• Gradual visualization (understanding): from the most important to the less
– Provision of top-k diagrams for successive values of k
top-1 top-3 top-5 top-8 Real diagram
• Other applications: Keyword searching, ...
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 73
Remark: Web Searching is an Analogous problem
• The Web graph is a very big graph too.
• Link Analysis has been proved very successful in Web searching
(and recently in many other domains)
– Main techniques: PageRank, HITS.
• Idea: Define a PageRank-like scoring scheme for ER diagrams
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 74
Applying PageRank on ER diagrams
Differences with the Web
• Directed links
• Binary links
• ignore self hyperlinks
Person CitylivesIn
fatherOf
Person should be ranked higher than City
• Undirected relationships
• n-ary relationships
• cyclic relationships are important
Web ER
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 75
Viewing an ER diagram as a Markov chain
• Entity type => state
• Binary relationship type => two counterpoising transitions
– An n-ary relationship type is first replaced by n(n-1)/2 binary relationship types
a
b
c
e
d
a
b
c
e
d
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 76
Towards defining EntityRank
Deriving the Probability Transition Matrix M
01100
12100
11012
00100
00200
e
d
c
b
a
edcba
A: Adjacency matrix
02/12/100
4/14/24/100
5/15/105/15/2
00100
00100
e
d
c
b
a
edcba
M: transition matrix
By normalizing each
row of A to sum to 1
a
b
c
e
d
1
2/5
1/5
1/5
1/4
1/2
2/5
1/41/5
1/2
1
a
b
c
e
d
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 77
What about ISA Hierarchies ?
• We could view an ISA link as transitions (one way or two-way).
• Our approach
– We ignore them, but
– We have an optional preprocessing step where we collapse each ISA hierarchy into one node (the root(s) of the hierarchy)
• that collects all attributes and relationship types
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 78
EntityRank
• Suppose a random E-R surfer who at each time step is at some entity type e and then:
– with probability q (e.g. q=0.15) jumps to a randomly picked entity type, and
– with probability (1-q) jumps to an entity type that is connected with the e
• EntityRank score := stationary probability
• This process defines a Markov chain with transition matrix:
– q • U + (1-q) • M
– where U[ei,ej]=1/N for all i,j, where N the number of entity types
– As the transition graph is strongly connected and non-bipartite, the fundamental
theorem of Markov chains apply.
– The score matrix is the principal right eigenvector of the following transition matrix:
(q • U + (1-q) • M)T
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 79
EntityRank (II)
• The matrix equation gives the following equation for each entitytype e:
∑∈
−+=)(' |)'(|
)'()1()(
econne econn
eScq
N
qeSc
conn(e): bag (duplicates allowed) with all entities types that are connected with e
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 80
B(iased)EntityRank
• The probability of jumping to a random etype is not the same for all, but it depends on the number of its attributes.
• This process defines a Markov chain with transition matrix:
– q • B + (1-q) • M where|attributes all|
|)(|],[B
j
ji
eattrsee =
∑∈
−+=)(' |)'(|
)'()1(
|allattrs|
|)(|)(
econne econn
eScq
eattrsqeSc
• BEntityRank is a well founded method for incorporating external
domain knowledge, and preferences:
– Number of tuples in the associated database tables
– Application programs’ calls
– User feedback while interacting with top-k diagrams
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 81
Evaluating Ranking Methods for Conceptual Diagrams
Evaluation is difficult.
The most safe evaluation is to test it on well known schemas
Kinds of Evaluation
• Empirical
• Formal (TREC/INEX-like)
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 82
Empirical Evaluation of Ranking Methods for Conceptual Diagrams
Typical Distribution of Scores
Large ER diagrams tend to have a well-connected kernelZiph’s Law
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 83
Emp. Evaluation on DICO1
Top-3 diagram of
0-N
0-N
SUIVI ACTIVITE
remplace
0-1
est remplace
0-1 Remplacement
0-N
0-N
POSITIONS/ACTIVITE/MOIS0-N
0-N
PAYS ENSEIGNE LANGUE
0-N
1-1description
0-1
dι crire
0-N
Notice
0-N
0-N
Mouvement de stock
0-N
0-N
FRAIS PAR DEFAUT
pays d'origine
0-N
pays destination
0-N
FRAIS DE DOUANE
Cadeau associι
0-N
0-N
EARLY BIRD
article vpc ensemble
0-Narticle vpc piθce
0-N
dι composition d'un ensemble
0-N
0-N
calcul goulotte
0-N
0-N
calcul canal
0-N
0-N
ALIMENTER
SECTEUR ACTIVITE
PAYS
ARTICLE VPC
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 84
Top-5 diagram of
0-N
1-N
SUIVI OPERATRICE
0-N
0-N
0-N
SUIVI ACTIVITE
remplace
0-1
est remplace
0-1
Remplacement
0-N
0-N
POSITIONS/ACTIVITE/MOIS
0-N
0-N
Poids max/canal
0-N0-NPAYS ENSEIGNE LANGUE
0-N
1-1description
0-1 dι crire
0-N
Notice
0-N
0-NMouvement de stock
0-N 0-N
FRAIS PAR DEFAUT
pays d'origine0-N
pays destination
0-N
FRAIS DE DOUANE
Cadeau associι
0-N
0-N
0-N
EARLY BIRD
article vpc ensemble
0-Narticle vpc piθce
0-N
dι composition d'un ensemble
0-N
0-N
cumul mouvement systeme mesure
0-N
1-1
cn/pay
0-N
0-N
0-N
calcul goulotte
0-N
0-N
0-N
calcul canal
0-N
0-N0-N
ALIMENTER
SECTEUR ACTIVITEPAYS
ETAT D'AVANCEMENT (LIGNE OU CO
CANAL
ARTICLE VPC
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 85
Top-11 diagram of
0-N
0-N
SUIVRE LE DOSSIER
1-N
0-N
SUIVI OPERATRICE
0-N
0-N
SUIVI LIVRAISON COLIS
0-N
0-N
SUIVI DES TOPS CLIENT
0-N
0-N
1-1
SUIVI COMMANDES
0-N
0-N
0-N
SUIVI ACTIVITE
1-N
1-N
STAT FOURNISSEUR MENSUELLES
0-N
0-N
simulation prix de vente
0-N
0-N
rι expι dition colis
est remplace
0-1 remplace
0-1
Remplacement
date dι but de validitι
0-N
date de fin de validitι
0-N
0-N
remise fin annι e fournisseur
0-N
0-N
POSITIONS/ACTIVITE/MOIS
0-N
0-N
POS-DTE
0-N
1-N
pondι ration
0-N
0-N
Poids max/canal
0-N
1-N
PAYS LANGUE
0-N
0-N
0-N
PAYS ENSEIGNE LANGUE
0-N
1-1
pays du fournisseur
0-N
1-1
pays de la devise
0-1
0-N
Pays d'origine
dι crire
0-N
description
0-11-1
0-N
0-N
Notice
0-N
0-N
Mouvement de stock
0-N0-N
Date de depot
1-1LOT
1-10-NLIVRAISON COLIS
1-N
1-1
HISTORIQUE COLIS
0-N0-NFRAIS PAR DEFAUT
pays destination
0-N
pays d'origine
0-NFRAIS DE DOUANE
1-1
0-N
fournisseur/pays
1-1
0-N
fournisseur/langue
0-N
0-N
FORMULE PAR PAYS/LANGUE
0-N
0-N
0-N
Cadeau associι
0-N EARLY BIRD
article vpc piθce
0-Narticle vpc ensemble
0-N
dι composition d'un ensemble
0-N
0-N
dte/pay
1-1
0-N
Document par langue
0-N
0-N
cumul mouvement systeme mesure
0-N
0-N
Composition
0-N
0-N
colcom/suicol
0-N
1-1
col/eta
1-10-N
cn/suicol
1-1
0-N
cn/pay
1-1
0-N
cd/suicol
0-1
0-N
cd/eta
1-1
0-N
Canal livraison
1-1
0-N
Canal commande
0-N
0-N
0-N
calcul goulotte0-N
0-N
0-N
calcul canal
1-1 0-Natvsec/cd
0-N
0-1
artvpc/frn
0-N
1-N
artref/tva
0-N
1-N
artref/douref
1-N
0-N
artref/atvsec
0-N
1-1
artref/artvpc0-N
0-N
0-N
0-N
ALIMENTER
1-1
0-N
achfrn/lan
SECTEUR ACTIVITE
REFERENCE ARTICLE
PAYS
LANGUE
FOURNISSEUR
ETAT D'AVANCEMENT (LIGNE OU CO
DATE
COMMANDE
COLIS
CANAL
ARTICLE VPC
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 86
Emp. Evaluation on the TELEBIB schema
NO TE: Th e re latio ns hip s in blu e are g eneric relat ion shi ps t ha t are de taile d b y s peci fic relat ion sh ips.
0-1
0-N
Wo rk ersCo mp ensa tio nClaim-Polic yHo ld er
1-1
0-NWorkers Co mp en satio nCla im-N ot ifier
1-1
0-N
Wat erVeh icle-Sea tin gCap acit y
1-1
0 -NWaterVehi cle-Loa dCap acit y
1-1
0-NWat erVe hicle -Eng ine
1-1
0-N
Victim-Pro fes sion Se nio rit y
1-1
0 -1
Vict im-Mut ua lHealt hIns urer
1 -1
0-N
Vict im-Job Ca teg orySen iority
1-1
0-N
Victim-Co mpa ny Sen io rit y
1-1
0-N
Vict im-Branc hSeni ority
1-1
0 -N
Vict im-Bon usRec ept ion Decl arat i
0 -N
1-1
Vic tim-AIKRece pt ion Decla rat ion
1-10-1
Vict im-A ct ualIns urer
1-1
0 -N
Victim-A cti vit yExe cut ion A dd res
1-1
0-1VERIDA SSReg istratio n-Req uest er
0-1
0-N
Vehic le-Trans po rt edGo od s
0-1
0-N
Ve hicl e-Secu rit ySyst em
1-1 0-1Veh icle-Regis tratio nPlat eHo lde
1-1
0-1
Veh icle-Mo to rTPL
1-1
0-1
Veh icle-Mo to rTh eft
1 -10-1Veh icle-Mot orPa rt ialComprehen siv e
1-1
0-1
Ve hicle -Mo to rFul lComp reh en siv e
1-1
0 -1
Veh icle-Mo to rFire
1-1 0-1Veh icle-Lea sing Compan y
1 -1
0-N
To talLo ss-Req u estForBid
0-1
0-N
Sub gu arant ee-Co v er
1-1
0-N
Sp ecialCo nd iti on -Clau se
0-1
0 -NSett lemen t-Party
0-1
0-N
Sett lemen t-Guaran tee
0-1
0-N
Sett lemen t-Cov er
0-1
1-1 Risk Obje ct-Previo u sIn surer
1-1
0-N
Req ue stForBid-Bid/ Offer
0-1
0-NRe late dPart y-Rate
1-10-1 Rel ated Pa rty -O fficialA dd re ss
1-1
0-N
Re late dPart y-OfficeTele ph on e
1-1
0-N
Rel ated Pa rt y-M ob ileTele ph on e
1-1
0-N
Rel ated Pa rt y-H omeTelep ho ne
1 -1
0 -N
Relate dPart y-Fax
1-1
0-N Re lated Part y-Exte rn alRole
1-1
0-N
Rel ated Pa rty -e Mail
1-1
0-N Reg Leg Pe rso n-Party Role
1-1
0-1
Re gLegPers-Soci alSec urity Burea1-1
0-1
RegLeg Pers-Parity Committ ee
1-1
0-1
Re gLegPers-Ho lid ayPay Bureau
1-1
0-1
Reg Leg Pe rs-FamilyA llo wan ceBu reau
1-1
0-1Pro perty -Risk Ad dress
0-1
1 -1
Professio n -A ctiv ity Exec tu tio nA dd re ss
1-1 0-NPriva teVeh icle/ Tru ck-D riv er
1-1
0-N
PreNot -Ren ewa lPremiu mPerio d
1-1
0-N
PreNo t-Proratio nPre miumPeriod
0-N
0-1
0-N
PremiumNot ific atio n-PIRelat edA
1-1
0-1
Premiu mLine-Po licy Ho lder
1 -1
0-1
Premiu m-Tax
1-1
0-1
Premium-Sub scri pt ion In dex
1-1
0-1
Premiu m-Sp litPayment Lo ad ing
1-1
0-1
Pre mium-Prev iou sInd ex
1-1
0-1Premium-Ot herIn suran ceRed uct io
1-1
0-1
Premiu m-Fid elit yBo nu s
1-1
0-1
Premi um-Commerc ialDisc ou nt
1-1
0-1
Pre mium-A ctu alInd ex
1-1
0-N
Po licy Ho lder-RSRRe gist ra tio n
0-N
0-1
0-N
Policy Ho lde r-Pre vio us Po licy
0-N
0-1
Pol icyH old er-Previ ou sClaim
0 -1
0 -N
Poli cyH old er-PrevIns Re fu sal
0-N
1-1
Po licyH old er-Prev ent ion Mea sure
0-1
0-N
Po licy Ho lder-O ffence Pen din g
0-1
0-N
Pol icyH old er-NCBLe vel
0-1
0-N
Po licy Ho lder-Co nv ict ion
0-1
0-1
Policy Ho lde r-Bill ing Ad dress
0-1
0 -N
Polic yHo ld er-Acc iden ts Sin ceA dju stment
1-1
0-N
Phy sica lPerson -Party Ro le
1-1
0-N
Phy sica lIn jury-Treat ing Med ical
1-1
0-N
Ph ys icalIn ju ry -Ho sp ital
1-1
0-N
Ph ys icalInju ry -First A idSup plie
1-1
0-N
Phy sicalInj ury-Disab ility
1-1
0-N
Phy sicalInju ry -Certifyi ng Med ic
1-1
0-1
Part y-Previo us Ad dress
1-1
0-1
Pa rty -O fficialA dd re ss
1 -1
0-N
Party -Offic eTelep ho ne
1-1
0 -N
Party -M ob ileTelep ho ne
1-1
0-1
Party -M ailPrimaryA d dress
1 -1
0-N
Pa rt y-HomeTelep ho ne
1 -1
0-N
Part y-Fax
1 -1
0-N
Part y-eMail
1-10 -N Op po ne nt Th irdPart y-Circu mst anc eColu mn
1 -1
0 -N N on RegLeg Person -Party Role
1 -1
0-1
M oto ri sedRo ad Ve hicl e-Veh iclePasse ng er
1 -1
0 -N
Mo to rise dRo adVeh icle-Tech nica rDa ta
1-1
0-1
M ot orised Road Vehic le-Seat ing Capa city
1-1
0-1
M ot orised Ro ad Vehic le-Load Cap acit y
1-1
0-N
Mo to ris edRo adVe hicle -Hau lag eTra iler
1-1
0-1
Mo to ri sedRo ad Ve hicle -Eng ine
1-1
0 -N
Mo to rised Roa dVeh icle-Carav an
1-1
0-N
M ot orIn suranc eCon tract-VERIDA S
1-1 Mat erialOb ject -Ow ner
1-1 Mat eri alOb ject-Fi nan ceCo mpa ny
1-1
0 -1M aterialDa mag e-Rep airer
1-1
0-1
M aterialDa mag e-Da mag edO bjec t
1-1
0-NMa terialDamage -A ssess or
1-1
0-N
LifeIn sCon tract -Referen cePremium
1-1
0-N
LifeInsCo nt rac t-LifeProfitSha rin g
1-1
0 -N
LifeIns Co nt ra ct-Las tPremiu m
1-1
0-N
LifeIn sCon tract -First Premiu m
1-1
0-N
LifeInsCo nt ract -D eceas eProfitSharin g
1 -1
0-1
Lega lPerso n-Head OfficeA dd re ss
1-1
0-1 Le galPerson -Branc hOffice Ad dres
1-1
1-1
KfkOp erat io n-Sen din gCo mp any
1-1
1-1
K fkOp eratio n-Receiv in gCo mpa ny
1-1
0-N
In suredPerson -Primary Occu pa tio n0 -1
0-1
Ins uredPers on -Permanen tD isab il
1-1
0-1
Insu red Perso n-Gene ralLife
1 -1
0-N
Ins ure dPers on -A dd itio na lOccu pa tio n1-1
0-N
In suredPerson -A dd iti on alA ctiv itie s
1-1
0-N
In su red -Circumsta nceCo lumn
1-1
0 -N
Ins ura nce Prov ider-FileA dminis t
1-1
Inc iden t-Witn ess
1-1
In ciden t-Op po ne ntTh ird Part y
1-1
Inc iden t-In terve nin gA ut ho rit y
1-1 Inc iden t-Ev en tA dd res s
1-1
0-N
H un tin g-Hu nt ing Lic ence
1 -10-N Gu aran tee -Sub gu ara nt ee
1-1
0-N
Gu aran te e-Spec ialCon d itio n
1 -1
0-1
Guarant ee-Prop ortio na lPremiu m
1 -1
0-1
Gu ara nt ee-O ldPremium
1-1
0 -1
Guaran tee-Net Premiu mNo nInd exed
1-1
0-1
Guarant ee-Net Premiu mIn de xe d
1-1
0-N
Gu ara nt ee-Ins Va lBased Ded uct ibl e
1-1
0 -1
Guaran tee-Gro ssPremi um
1-10-N
Gu aran te e-Gen era lCon dit ion
1 -1
0-N
Guarant ee-Exte nsio n
1-1
0 -1
Gu aran te e-Da mA mBased Ded uct ib le
1-1
0-N
Guaran tee-Daily A llow ance
0-1
0-N
Gu ara nt ee-Co ve r
1-1
0-N
Gu arant ee-Co st /Cha rg e
1-1
0-N
Gu aran tee -Co mmiss ion
1-1
0-N
Guaran tee-Capit al
1 -1
0-1
Gu arant ee-BasePremi um
1-1
0-1
Gu aran te e-A nn ua lPremiu mNo nInd e
1-1
0-1
Gua ran tee-A nn ualPre miumIn dexed
1-1
0-N
Gua ran tee -A nn ualInd emn ificatio n
1 -1
0-NGene ralLife -Ben efici ary
1-1
0-N
Gen era lCon dit ion -Clau se
0 -1
1-N
Fle et-Vehic le
1-1
0 -N
Factu alGro u p-Party Role
1-1
1-2
EAN Fo rm-Circu mst an ceColu mn
0-10-N Drive r-Previo us Cl aim
0-1
0 -N
D riv er-NCBLe vel
1-1
0-N
D riv er-Driv in gLicen ceWith drawa
1 -1
0-1
Driver-D riv ing Lic enc e
0-1
0-N
D riv er-Ac cide nt sSin ceA dju st men t
1-1
0 -N
Do cu men t-D ocu men tExcha ng eDet ails
1-10-N D eclarati on -Resp on se
0-1
0-ND amag edO bje ct-Va luat io n
0-1
1-1
D amag e-Gu ara nt ee
1-1
0-N
Cu rA cc-Premiu mRetu rn Li ne
1-1
0-N
Cu rA cc-Premiu mReimb ursement Lin e
1 -1
0 -N
CurAcc -Premiu mInv oic eLin e
1 -1
0-N
CurA cc-PremiumCance llatio n Lin e
1-1
0 -N
Cu rA cc-Misc ellTran sLine
1 -1
0-N
Cu rA cc-Fin Tran sLine
1-1
0-N
CurA cc-Co mmis sion Li ne
1-1
0-N
Cu rA cc-ClaimLin e
1-10-N CurAcc -Bo rd ereau Tot alLine
1-1
0 -1
Co v erLimit -Sub scrip ti on In dex
1-1
0-1
Cov erLimit -Prev io usInd ex
1-1
0-1
Cov erLimit -A ctu alInd ex
1-1
0-N
Co up le-Party Ro le
1-1
0 -1
Con tract-Tran sfer
1-1
0-1
Co n tra ct-Su spe nsio n
1-1
0-N
Co nt ract -Spe cialCo nd itio n
0-1
0 -N
Con tract -Relat edPo licy
1-1
0-1
Con tract-Prop ortio n alPremiu m0-1
1 -N
Con tract -Premi umN ot ifica tio n
0 -1
1-N
Co nt ra ct-Po licy Hold er
1-1
0-1
Co nt rac t-Po licyCh an ge
1 -1
0-1
Con tract-Net Premiu mNo nInd exed
1-1
0-1
Con tract-Net Premiu mInd exed
1-N
0-N
Co nt rac t-Int ermedi ary
1-N
0-N
Co n tra ct-In surer
1-N Con tract-In su ran ceOb ject
1-1
0-1
Co nt rac t-Gross Premiu m
1 -1
0-N
Con tract-Gene ralCo nd itio n
1-1
0-N
Co n tra ct-Cost /Ch arg e
0-N
1-1
Co n tra ct-CoPolic yHo lde r
1 -1
0-N
Co nt ra ct-Co n tra ctD ocu men t
1-1
0-N
Con tract -Co mmiss ion
1-1
0-N
Co nt ract -Co ins urer
0 -1
0-N
Con tract -Claim
1 -1
0-1
Co nt ract -Can cella tio n
1-1
0-1
Con tract-An n ulme nt
1-10-1 Co nt ract -A nn u alPremiu mNo nInd ex
1-1
0-1
Con tract-An n ualPremiumIn dexed
1-1
0 -1
Co nt rac t-A me nd men t
0 -1
0-N
Co nt en ts-Secu ri ty Sy stem
0-1
0-1
Co n ten ts-Gene ralFire
0 -1
0-1
Co nt ent s-Bu rg lary/ Theft
1 -1
0-1
ClaimLin e-Pol icyH old er
0-1
0 -N
Cl aiman t-Set tle ment
0-N
Claiman t-Professio n
0-10-N Cla iman t-Damag e
0-N
0-1
Cla im-M ana gement Act
1-1
0-N
Cl aim-K fkO perati on
1-1
0-N
Claim-In termed iary
1-1
0-1
1-N
Cl aim-Cla iman t
1-1
1 -6
Circu mst an ceColu mn -Circumstan ce
0-1 0 -NBuil din g-TakeOv erPo licy
0-1
0-1
Bu ild ing -Sto rmH ailA nd Heav ySn ow
0-1
0-N
Bu ildin g-Se curity Sy ste m
0-1
1-1
Bu ild ing -Poli cyH old er
0-1
0-1 Build ing -Ge neralFire
1 -1
1-N
Bordereau -Borde reau Line
1 -1
0-N
Bon us Recep tio nDe claratio n-Bon us
0-1 As sessment -Sett lemen t
1-1
A ssessment -D amag edO bjec t
1-1
0 -N
AirVehic le-Seat ing Capa city
1-1
0-N
AirVeh icle-Lo adCap acit y
1-1 0-NAirVehi cle-En gin e
1 -1
0-N
A IKRec ept ion Dec larat io n-A IK
0-1
0-1
A ccou nt in gDo cu men t-Int ermed iar
0-1
0-1A cco u nt ing Doc ume nt -In su rer
P
WorkersComp en sati on Con tra ct
Wo rk ersCo mp ens atio nCla im
Wo rkersCo mpen sat ion
Wit nes s
WaterVeh icle
Vi ctim
VERIDA SSReg istrat ion
Veh icle Re gist rat io nPlate Hold er
Ve hicle Regist ra tio nCard
Vehi clePassen gers
Veh icle
Va luat io n
Trea tin gM edic alDo cto r
Trav elIn su ran ceCon tract
Trans po rtIns uranc eCon tract
Tran spo rt edGo od s
Tran sfer
To talTe mpo raryD isab ility
To talLo ss
ThirdCo mmis sion
Temp orary Disa bilit y
Te leph o ne
Tec hn icarData
Tax
Ta riffPremium
Ta keOv erPol icy
Sus pen sio n
Su bsc rip tio nInd ex
Su bg ua ran tee
Sto rmHailA nd Hea vy Sn ow
St art In dex
Spl itPremi umPay men tLoad ing
Specia lCon dit ion
Spac eVehic le
Soc ialSecurity Bureau
Set tlement
Sen iority
Se nd ing Compan y
Sec urity Sy stemCert ific ate
Se curity Sy st em
Seco nd Re side ncy A dd ress
Sec on dCo mmiss ion
Seati ng Capac ity
RSRRegi stratio n
Roa dVeh icle
Ris kOb ject
RiskA dd re ss
Resp on se
Resid ent ialCarav an
Req ues tForBi d
Requ est er
Represen tat ive
Rep laced Po licy
Rep aire r
Ren ewa l
Relat edPol icy
Relat edParty
Re gist eredLega lPerso n
Refere ncePre miu m
Receiv in gCompan y
RealEstat e
Rate
Rai lVehic le
Prorat ion
Prop ortio nal Pre miu m
Pro pe rty
ProfitSharing
Professio nSen iority
Priv ateM ot orised Road Vehic le
Priv at eBuild ing
Prima ry Occu pa tio n
Previ ou sPo licy
Prev iou sIn su rer
Prev iou sIn su ran ceRefusal
Prev iou sIn de x
Pre vio usCla im
Prev io usA dd re ss
Preven ti on Measu re
Premiu mRetu rn Li ne
PremiumRe imb ursemen tLine
Pre miumPeriod
PremiumN ot ific atio n
Pre miu mLo ad ing
PremiumLin e
PremiumIn vo iceRelat ed Act ion
Premiu mIn vo iceLine
Premiu mIn v oice
Pre miumDisco un tO rLoad in g
Premiu mDi scou nt
Pre miumCance llatio n Lin e
Pre miumAn dCo mmis sion Rev ersib lePremiumLin e
Premiu m
Po licy Ho lder
Pol icy
Phy sical Person
Phy sicalInju ry
Perso n alLia bilit yIns urance Con tra ct
Permane nt Disab ilit yCov er
Permane nt Disab ility
Party Role
Part y
Partia lTempo ra ry Disab ilit y
Parity Committ ee
Own er
Ot he rPriv ateBu ild ingO th erPoli cy
Oth erIn suranc eRedu ct ion
Op p on ent ThirdPart y
Old Premiu m
O fficialA dd res s
OfficeTel eph on e
Offen cePend ing
Not ifie r
Non Reg istered Leg alPerso n
No nM ot orised Ro ad Vehic le
N etPremiumNon in dexed
N etPremiumIn dexed
N etPremium
NCBYearMin us3
NCBYearM inu s2
NCBReferen ce
NCBPrev iou sNCBLevel
NCBAt Un derwri tin g
NCBA tPrev io usInsu re r
N CBA ct ual
M ut ualH ealt hInsu rer
Mo to rTh ird Pa rty Liabil ity
M ot orTheft
M ot orPartialCo mp reh en sive
Mo to ris edRo adVeh icle
M ot orIn suran ceCon tract
Mo to rFullCo mp reh en sive
M ot orFire
M ot orCycleCe rti ficat e
Mo to rCo mprehe nsiv e
Mo rt gag eLoan Fi le
MoralDa mag e
M ob ileTelep ho ne
Mis cellan eo usTrans actio nLin e
Me dica lDoc to r
M aterialDa mag e
Man ag emen tA ct
Ma ilSec on daryA d dress
MailPrimary Ad dress
M achin ery/ Ap plia nce s
Lo adCa pacit y
LifePro fi tSharing
LifeIn su ran ceCo nt ract
Le galPerson
Legal Expe nses In suran ceCon tract
Leas ing Compan y
Last Premiu m
Last Commissio n
Lan d
K fkReq ue stForRefu nd
KfkOp erat io n
KfkDeb itO rCre dit St atement
Jo bCat eg orySenio ri ty
Int erv enin g Au th ority
In te rmed iary
In su rer
In suredVa lueBas edD edu cti ble
In sured Perso n
Insu re d
In suranc eProv ide r
In div idu alInsu ra nceCo nt ra ct
In de x
In demn ificati on
Iden tit yCard
Hu nt ing Lic ence
Hu nt ing
Ho spi tal
HomeTelep ho ne
H olid ay Pay Bureau
He adO ffi ceA dd ress
H aula geTrailer
Gu aran tee
Gro up In su ran ceCo nt ract
Gro up
Gross Premiu m
Gree nCard
GeneralLife
Gene ralFire
Gene ralCo nd itio ns Wo rd ing
Gen eral Co nd it ion
Gen eral Ci vilLiab ility
Gara ge
Fu tu reA dd res s
Fleet
Fi rst Premiu m
Firs tCo mmissi on
First Aid Sup plier
FireIn su ran ceCo nt ract
Fin ancia lTran sact ion Li ne
Fin anc ialDamage
Fina nceCo mp any
Fi leAd min ist rat or
Fid elit yBo nu s
Fee
Fax
Family A llow anc eBureau
Fact ual Grou p
Exte rna lRole
Ext en sion
Exp ense
Ev ent A dd ress
Europ ean Ac cide nt Not ific atio nForm
En gin e
En dInd ex
Emp lo yee Be nefitsCo nt rac t
eMa il
Dw ellin g
Drivin g Licen ceWit hd raw al
Driv ing Licen ce
Drive r
D ocu men tExcha ng eDet ails
Do cument
Dis abili ty
Direct Pay men t
De du ctib le
Declarat ion
Decea seProfitSh arin g
Da taEnt ry Ope rat or
Dama ged Ob ject
Dama geA mo un tBase dDe du ctib le
Damage
Daily A llow ance
Curren tA cco un tLine
Cu rrent Ac cou nt
Cred ito r
Co verNo te
Co verLi mit
Co ver
Cou p le
Co st/ Charge
Co Policy Ho lde r
Con vict io n
Con tract Doc ume nt
Co nt rac tCha ng e
Co nt ract
Co nt ent s
Con di tio n
Compan ySen iority
Commun icat ion Cha nn el
Co mmo nLaw
Commissio nLine
Commissio n
Co mmercialM ot orised Road Vehi cle
Commerc ialLoss
Commerci alIn su ran ceCon tract
CommercialDis cou nt
Co mmercia lBuild ing
Coll ectio n
Co ins urer
Claus e
ClaimLin e
Cla iman t
Cl aim
Civ ilLiabili ty HorsesA nd Carriag es
Circu mst an ceColu mn
Circ ums tan ce
Certifyin gM edi calDo cto r
Caravan
Capit al
Ca ncel latio n
Calcu latio n
Burgla ry /Th eft
Build in g
Bra nch Se nio rit y
Bran chO ffi ceA dd ress
Bordereau Tot alLin e
Bo rd erea uLine
Bo rd ereau
Bo nu sRecep tio nD eclarati on
Bon us
Bil ling Ad dress
Bi d/ Offer
Bicy cle
Ben eficia ry
BasePre mium
As sesso r
An nu lmen t
A nn u alPremiu mNo nInd exed
A n nu alPremiu mInd exed
An nu alPremium
A nn u alIn demnificat ion
A men dme nt
Ai rVeh icle
A dv ant ag eIn Kin dRece pt ion Decla rat ion
A dv ant age In Kin d
A d dress
Ad dO nInsu re r
Ad di tio nalO ccu pat ion
A d dit ion alA ctiv ity
Act ua lIn surerO f
A ctu alInd ex
Act ivi ty Execu tio nA d dress
Act ion
Acq u isitio nCo mmi ssio n
A cco un tin gD ocu men tLine
A cco un tin gD ocu men t
Ac cide nt sSin ceA dj ust men t
Ove rvi ew/Co nc ept ual
Belgian national standard for
exchanging messages between
insurance companies
(339 entity types and
232 relationship types)
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 87
Emp. Evaluation on the TELEBIB schema
0-1
1-N
Contract-PolicyHolder
Victim
Vehicle
Premium
PolicyHolder
MaterialObject
Guarantee
CurrentAccount
Contract
NOTE: Th e re la ti o ns h ip s i n b lu e a re g e n e ric re la ti o ns h ip s t h at a re de t ai le d b y s p ec if ic r e la ti on s h ip s .
0- 1
0- N
W o rk er s Co mpe n s at io n Cla im -Po lic y Ho ld e r
1 - 1
0- NW o rk e rs Co mp en s a ti o nCl ai m-No t if ie r
1 - 1
0- N
Wa t er Veh i cl e- Se at in g Ca pa c it y
1- 1
0 - NWa t er Veh i cl e- Lo ad Ca p a ci ty
1 -1
0- NW at er Ve h ic le -En g in e
1 -1
0 -N
Vic ti m-Pr of e ss i on Se n io ri ty
1- 1
0 -1
Vi ct im- Mu t ua lH ea lt h In s u re r
1 -1
0 -N
Vic ti m-J o b Cat e go r y Se n io r it y
1- 1
0- N
Vic t im-Co mp a n y Sen io r it y1- 1
0 -N
Vic t im- Bra n ch Se n io ri ty
1 -1
0 -N
Vi ct im- Bo n us Re c ep t io n Dec l ar a ti
0 -N
1- 1
Vic ti m-AI KRe c ep t io n Dec l ar a tio n
1 -1 0- 1Vic t im- Ac tu a lI ns u re r
1 -1
0 - N
Vic ti m-Ac t iv it y Ex e cu t io n Ad d re s
1 -1
0 -1VERIDASSRe g is tr a tio n - Req u e st e r
0- 1
0- N
Ve h ic le -Tr a ns p o rt e dGo o d s
0 -1
0- N
Ve h ic le -Se c u rit y Sys t em
1 -1 0 -1Veh i cl e -Re g is tr at io n Pla te Ho ld e
1 -1
0 -1
Ve hi c le -M ot o rTPL
1 -1
0 - 1Ve hi cl e -M ot o rTh e ft
1 - 1 0 -1Ve h ic le - Mo to r Par ti al Co mpr e h en s iv e
1 - 1 0 -1Ve h ic le -M o to rFu l lCo mp re h en s iv e
1 - 1
0 -1
Ve hi c le -M ot o rFir e
1 - 1 0 -1Ve h ic le -Le a si n gCo mp a n y
1- 1
0 - N
To ta lLo s s -Re q u es tFo r Bid
0 -1
0- N
Su b g u a ra n te e -Co v er
1 -1
0- N
Sp e ci a lCo n di ti on - Cla u se
0 -1
0 - NSet tl e men t -Pa rt y
0- 1
0- N
Se tt le men t -Gu a ra n te e
0 -1
0 -N
Se t tle me n t- Co ve r
0 - 1
1 - 1 Ris k Ob je ct -Pr e vi o us I ns u r er
1 -1
0- N
Re q u e st For Bid - Bid /Of fe r
0 -1
0 - NRe la t ed Pa rt y- Ra te
1 -10- 1 Re l at ed Pa rt y -Of fi ci al Ad d re s s
1 - 1
0 -N
Re la te d Par ty - Off ic e Tel ep h o n e
1 -1
0 -N
Re la t ed Pa rt y- Mo b il e Tel ep h o n e
1- 1
0 -N
Re la t ed Pa rt y -Ho me Te le p ho n e
1- 1
0- N
Re la te d Par ty - Fa x
1- 1
0 - N Re la te d Pa rt y- Ext e rn a lRo le
1- 1
0 - N
Re la te d Pa rt y- e Ma il
1 -1
0 - N Reg Le g Pe r so n - Pa r ty Ro le
1 -1
0 -1
Re g Le gPe rs - Soc ia l Se c u ri ty Bu re a1- 1
0- 1
Reg Le g Per s- Pa rit y Co mmit te e
1 -1
0 - 1
Re g Le g Pe rs -Ho l id a yPa y Bu re a u
1 -1
0 -1
Re g Leg Pe rs - Fa mi ly All o wan c e Bu re a u
1 - 1
0- 1Pr o pe r ty -Ri s kAd d r es s
0 -1
1- 1
Pro f e ss io n - Ac tiv i ty Exe c tu t io n Ad d re s s
1 -1 0 - NPr iv a te Ve hi c le /Tr uc k -Dr iv e r
1 - 1
0 -N
Pr eNo t -Re n e wal Pre miu mPe ri od1- 1
0 - N
Pre No t- Pro r at io n Pre miu mPe ri o d
0- N
0 -1
0 - N
Pre miu mNo t if ic at io n -PI Rel a te d A
1 -1
0 -1
Pre mi um Li n e- Po lic y Ho ld e r
1 -1
0 -1 Pre miu m- Ta x
1 - 1
0 -1
Pre miu m- Su bs c ri p tio n I nd e x
1 -1
0 -1
Pre miu m- Spl it Pay me n tLo a d in g
1 -1
0 -1
Pre miu m- Pre v io u sI n d ex
1- 1
0 -1Pre miu m- Oth e rI n su r an c e Red u c ti o
1- 1
0 -1
Pre miu m- Fid e lit y Bon u s
1 - 1
0 -1Pre miu m- Co mmer c ia lDi sc o u n t
1- 1
0- 1
Pre mi u m-Ac t ua lI n d ex
1 -1
0 -N
Pol ic y Ho ld e r- RSRRe g is t ra ti o n
0 -N
0 -1
0 -N
Po lic y Ho ld e r- Pre v io u sPo l ic y
0 - N
0 -1
Pol ic y Ho ld e r- Pr e vi o u sCl ai m
0 -1
0 -N
Po l ic y Ho ld er -Pr e vI n sRe f us a l
0- N
1 - 1
Po li c yH ol d er -Pr ev e n ti on M e as u re
0 -1
0 -N
Po li c yHo l d er -Of fe n c ePe n d in g
0 -1
0- N
Pol ic y Ho ld e r- NCBLev e l
0 -1
0 -N
Pol ic y Ho ld e r- Con v ic t io n
0 -1
0 -1
Po lic y Ho ld e r- Bill in g Ad d re s s
0- 1
0 - N
Pol ic y Ho ld e r- Ac ci d en t sSi nc e Ad ju s tme n t
1 - 1
0- N
Ph y si ca lPe r so n -Pa rt y Ro le
1 -1
0 -N
Phy s ic a lI n ju ry - Tre a tin g M ed i ca l
1 -1
0 -N
Phy s ic a lI n ju ry - Ho sp it a l
1 -1
0 -N
Ph y s ic a lIn j ur y -Fir s tAi d Su p p li e
1 - 1
0 -N
Ph y s ic al In j ur y -Di sa b il it y
1- 1
0 -N
Ph ys i ca lI n ju ry - Cer ti fy in g M ed i c
1 -1
0 -1
Par ty - Pre v io u sA d dr e ss
1 -1
0 - 1
Par ty - Off ic ia lA dd r es s
1 -1
0- N
Par ty - Off ic e Tel ep h o n e
1 -1
0 -N
Par ty - Mo b ile Te le p h o ne
1- 1
0 -1
Par ty - Ma il Prima r y Ad dr e ss
1 -1
0- N
Par ty - Ho meTe le p h o n e
1 -1
0- N
Par ty - Fax
1 -1
0- N
Par ty - eM a il
1 -10 -N O pp o n e n tTh ir d Pa r ty -Ci rc u ms ta n ce Co lu mn
1 - 1
0 - N No n Re gLe g Per s on - Par ty Ro le
1- 1
0- 1
M ot o ri se d Ro ad Ve h ic le - Veh ic l ePa s se n g er
1 -1
0 -N
Mo t o ris e d Ro ad Ve h ic le -Te c h ni c ar Da ta
1- 1
0- 1
Mo t o ris e d Ro ad Ve h ic le -Se a ti ng Ca p a ci ty
1 - 1
0 - 1
M ot o ri se d Ro ad Ve h ic le - Lo a d Ca pa c it y
1 -1
0 -N
M o to r is ed Ro a d Veh i cl e- Ha u la ge Tr ai le r
1 -1
0 -1
Mo to r is e dRo a d Ve hi cl e -En g in e
1 -1
0 -N
Mo to r is e dRo a d Ve hi cl e -Ca ra v a n
1 - 1
0 -N
Mo t or In s u ra n c eCo n tr a ct -VERI DAS
1 -1 Ma t er ia lOb j ec t- Own e r
1 -1 Ma t er ia lOb j ec t -Fin a n ce Co mp a ny
1 -1
0 - 1M at e ria l Dama g e -Re p a ire r
1 -1
0 -1
M at e ria l Dama g e -Da ma g e dOb j ec t
1 -1
0- NM a te ri al Da mag e -A ss e s so r
1 -1
0- N
Lif eI n sCo n t ra ct -Re f er e nc e Pre miu m
1 -1
0 -N
Lif e In s Co nt ra c t- Life Pr of it Sha r in g
1 -1
0- N
Lif e In s Con t ra c t- Las tPr e miu m
1- 1
0 -N
Life I ns Co n tr a ct -Fir s tPre mi u m
1 -1
0 - N
Li fe In s Co n tra c t- De ce a s ePr o fit Sh ar in g
1 - 1
0 -1
Leg a lPe rs o n -H ea d Off ic e Ad d re s s
1- 1
0 -1 Le g a lPe rs o n -Br an c h Off ic e Ad d re s
1 - 1
1 -1
K fk Op er a ti on - Sen d in g Co mp a ny
1 -1
1 - 1
Kfk Op e ra ti o n- Re ce iv i n gCo mp a n y
1- 1
0- N
In s ur e d Pe r so n -Pr ima r yOc c u p at io n0 -1
0 -1
In s ur e d Pe r so n -Pe r man e n tDi sa b il
1- 1
0 -1
I n s ur e dPe r so n -Ge n e ra lLi fe
1 -1
0 - N
In s u re d Per s on - Ad d it io n al Oc cu p a ti on 1 -1
0 -N
In s ur e d Pe r so n -A d di ti on a lA ct iv it ie s
1 -1
0- N
I n su r ed - Cir cu ms t an c e Col u mn
1 -1
0 -N
I ns u r an c e Pr o v id e r- Fi le Ad min i st
1 -1
I n ci d en t -W it ne s s
1 - 1
In c id e nt -O pp o n e n tTh ir d Pa r ty
1 - 1
I n c id en t -I n te rv e n in g Au th o r ity
1 -1 I n ci d en t -Ev e nt Ad d re s s
1 -1
0 -N
Hu n t in g -Hu n ti n gLi ce n c e
1 - 10- N Gu a ra n te e -Su b g ua r an t ee
1 -1
0 -N
Gu a ra n te e -Sp ec i al Con d i tio n
1 - 1
0 -1
Gu a ra n te e -Pr op o r tio n a lPr emi u m
1 -1
0 -1
Gu ar a n te e- Old Pr emi u m
1 -1
0 -1
Gu a ra n te e -Ne t Pre miu mNo n In d e xe d
1 -1
0- 1
Gu a ra n te e -Ne t Pre miu mI n de x ed
1- 1
0 -N
Gu a ra n te e -I n sVa lBa s e dD ed u c ti bl e
1- 1
0 -1
Gu a ra n te e -Gr os s Pre miu m
1 - 1 0 - N
Gu a ra n te e -Ge n er a lCo n di ti on
1 -1
0 -N
Gu a ra n te e -Ex te n s io n
1- 1
0 -1
Gu a ra n te e -Da mAmBa s e dD ed u c ti bl e
1 -1
0- N
Gu ar a nt e e- Da ily A llo wa n ce
0 -1
0- N
Gu a r an t ee - Co v e r
1 -1
0 -N
Gu ar an t e e- Co st /Ch a rg e
1 - 1
0 -N
Gua r an t ee - Com mis si o n
1 -1
0 - N
Gu a ra n te e -Ca p it al
1 - 1
0 -1
Gu ar a n te e- Ba se Pre mi um
1 - 1
0- 1
Gu a ra n te e -An n u a lPre mi um Non I n d e
1 -1
0 -1
Gu ar a nt e e- An n u al Pre miu mI n de x ed
1 - 1
0 -NGu ar a nt e e- An n u al In d e mn ifi c at io n
1- 1
0 - NGe n er al Lif e- Ben e f ic ia ry
1- 1
0 -N
Ge n e ra lCo n d it io n -Cla u s e
0 -1
1 -N
Fl ee t -Ve h ic le
1 -1
0 -N
Fa c tu a lGr ou p - Pa r ty Ro le
1 -1
1 -2
EANFo rm- Cir cu ms t an c e Col u mn
0 -10- N D riv e r- Pre v io u s Cla im
0- 1
0 - N
Dr iv e r- NCBLe v e l
1 - 1
0 - N
Dri ve r -Dr iv in g Lic e n ce W ith d r awa
1 -1
0 -1
Dri v er -Dr iv i ng Lic e n ce
0- 1
0 -N
Dr iv e r- Ac ci d en t sSi nc e Ad ju s tme n t
1 -1
0 - N
Do c u me nt -D oc u me n tEx ch a n g eDe t ai ls
1- 10- N De c la ra ti o n- Re sp o n s e
0 - 1
0- NDa mag e d Ob je c t- Val u at io n
0 -1
1 -1
Dam ag e -Gu a ra n te e
1 -1
0 - N
Cu rAc c -Pr em iu mRe tu rn Lin e
1 -1
0 -N
Cu rA cc - Pr e miu mRe imb u rs e me n tLin e
1 -1
0 -N
Cu r Acc - Pre miu mI nv o i ce Lin e
1 -1
0 - N
Cu rAc c -Pr emi u mCan c e lla t io n Lin e
1 - 1
0 - N Cur Ac c -Mi s ce ll Tra n s Li n e
1 -1
0- N
Cu r Acc - Fin Tra n s Li n e
1- 1
0 -N
Cu r Ac c- Co mmis s io n Lin e
1- 1
0- N
Cur Ac c -Cla i mLin e
1 -10- N Cu r Ac c- Bo rd e re a u To t a lLin e
1 -1
0 -1
Co v e rLimi t- Sub s c ri pt io n In d e x
1 -1
0 -1
Co v er Limi t- Pr e v io u sI n d ex
1- 1
0 -1
Co v er Limit -A ct u al In d e x
1 -1
0 -N
Co u p le -Pa r ty Ro le
1- 1
0 -1
Co n tr ac t- Tra n s fe r
1 -1
0 -1
Con t ra c t- Sus p e n si on
1 - 1
0 -N
Co n tr ac t- Sp ec i al Con d it io n
0 -1
0 -N
Co n tr ac t -Re la te d Po lic y
1 -1
0- 1
Con t ra c t- Pro p o rt io n al Pre miu m0- 1
1- N
Co n tr ac t -Pre mi um No tif ic a ti on
0 -1
1 -N
Co n tr a ct -Po li c yHo l d er
1 -1
0- 1
Co n tr a ct -Po li cy Ch a n g e
1 - 1
0 - 1
Co n tr a ct -Ne t Pr e miu mNo n In d e xe d
1- 1
0 - 1
Con t ra c t- Net Pre miu mI n d ex e d
1 - N
0- N
Co n tr a ct -I n te rme d ia r y
1 -N
0 -N
Con t ra c t- In s u re r
1- N Co n tr ac t- In s u ra n c eO bj ec t
1 -1
0- 1
Co n t ra c t-Gr o s sPr emi u m
1 - 1
0 -N
Co nt ra c t- Gen e r al Con d i tio n
1 -1
0 -N
Co n t ra ct -Co s t/ Ch a rg e
0 - N
1 -1
Co n tr a ct -Co Po lic y Ho ld e r
1 -1
0 -N
Con t ra c t- Con t ra c tDo c u me nt
1 -1
0 -N
Co n tra c t- Co mmis s io n
1 -1
0 -N
Co n t ra c t-Co i n su r er
0 - 1
0- N
Co n tr a ct -Cl ai m
1 -1
0 -1
Co n tr ac t -Ca n ce ll a tio n
1 -1
0 -1
Co n tr ac t -An n u lme n t
1 -10 -1 Co n t ra c t- An nu a lPr em iu mNo n In d e x
1 -1
0- 1 Co n tr a ct -A nn u a lPr emi u mIn d e xe d
1- 1
0- 1
Co nt ra c t- Ame n d men t
0 -1
0- N
Co n te n ts -Se c ur it y Sy s t em
0 -1
0 -1
Co n te n ts -Ge n e ra lFir e
0 -1
0 -1
Co n t en t s- Bu rg la r y/ Th ef t
1 -1
0- 1
Cl a imLin e -Po li cy Ho l de r
0 -1
0- N
Cl ai ma nt -Se tt le me n t
0 -N
Cla ima n t- Pro f es s io n
0 -10 -N Cl ai ma n t-D ama g e
0- N
0 - 1
Cla im- Ma n ag e me n tAc t
1 -1
0 -N
Cl ai m-Kf k Ope r at io n
1 -1
0- N
Cl ai m-I n te rm ed ia r y
1 -1
0 -1
1- N
Cla im -Cla im an t
1 - 1
1 - 6
Cir c ums t an c e Co lu mn -Ci rc u mst a nc e
0 -1 0 - NBu il d in g -Ta k eOv e rPo l ic y
0- 1
0- 1
Bui ld in g -St o rmHa il An d He av y Sn o w
0 -1
0 -N
Bu il di n g -Se cu r ity Sy s te m
0 -1
1 -1
Bu il d in g -Po li cy Ho ld e r
0 -1
0 -1 Bu i ld in g -Ge n e ra lFir e
1- 1
1- N
Bo r d er ea u - Bo r d er e au Lin e
1- 1
0 - N
Bo n u sRe c ep t io n De cl ar a ti on - Bo nu s
0 -1 As s es s me nt -Se tt le me n t
1 -1
A ss e s sme n t- Da ma ge d Ob je c t
1 -1
0 -N
A irVe h ic le - Se a ti n g Ca p a c ity
1 - 1
0 -N
Ai rVe h ic le - Lo a d Ca pa c it y
1 -1 0- NAi rVe h ic le -En g in e
1 - 1
0 -N
AIKRe c ep t io n De cl a ra ti on - AIK
0- 1
0- 1
Ac c o u nt in g Do c um en t -I nt e rme d ia r
0 - 1
0 - 1Ac co u n ti n gD oc u me n t- In s u re r
P
Wo r ke r sCo mp e n sa t io n Co nt ra c t
W o rk e rs Co mp en s a ti o nCl ai m
W o rk e rs Co mp en s a ti on
W it n es s
Wa t er Veh i cl e
Vic t im
VERI DASSRe gi st ra t io n
Ve h ic le Re g is tr at io n Pla te Ho ld e r
Veh ic l eRe g is tr a tio n Ca rd
Ve h ic le Pa ss e n ge r s
Ve hi c le
Va lu a ti o n
Tr ea ti n g Me d ic al Do ct o r
Tra v e lIn s u ra n c eCo n t ra ct
Tr an s p o rt In s u ra n ce Co n tr a ct
Tra n s p or te d Go o d s
Tra n s fe r
To ta lTe mp o ra r y Dis ab i lit y
To ta lLo s s
Th ird Co mmi ss io n
Temp o r ar y Dis a bi li ty
Te le p h o n e
Tec h n ic a rDa ta
Ta x
Ta rif fPr em iu m
Tak e Ov e rPo li cy
Su s p en s io n
Su b sc r ip ti o nI n d ex
Su b gu a ra n t ee
St o rmHa il An d He av y Sn o w
Sta r tI nd e x
Sp li tPr emi u mPay me n tLo a d in g
Spe c ia lCo n d it io n
Sp a c eVe h ic le
Soc i al Sec u ri ty Bu re a u
Se tt le me n t
Sen io r it y
Se n d in g Comp a n y
Se c u rit y Sy st em Cer ti fic a te
Se c u rit y Sys t em
Se co n d Re si d en c y Ad d re s s
Sec o n d Comm is si o n
Se a ti ng Ca p ac i ty
RSRRe gi st ra t io n
Ro a d Veh ic l e
Ri sk Ob je c t
Ri sk Ad d re s s
Re sp o n s e
Res id e n ti al Ca ra v an
Re q ue s tFo r Bid
Req u e st e r
Rep r es e n ta ti v e
Re p la c ed Po li cy
Rep a ir e r
Ren e wa l
Re la t ed Po li cy
Re l at ed Pa rt y
Reg i st er e d Le g a lPe rs o n
Re fe r en c e Pre miu m
Rec e iv in g Co mp a ny
Rea lEs t at e
Rat e
Ra il Ve h ic l e
Pro r at io n
Pro p o rt io n a lPre miu m
Pro p e rt y
Pro f it Sha r in g
Pro fe s s io n Sen i or it y
Pri v at e Mo to r is e dRo a d Ve hi c le
Pr iv a te Bu il di n g
Prima r y Occ u p a tio n
Pre v io u sPo li c y
Pre v io u s In s ur e r
Pre v io u s In s u ra n ce Re fu s a l
Pre v io u sI n d ex
Pre v io u s Cla im
Pre v io u s Ad d re s s
Pre v en t io n Me a su r e
Pre miu mRe t ur n Lin e
Pr e miu mRe imb u rs e me nt Lin e
Pre mi um Pe r io d
Pre miu mNo t if ic at io n
Pre mi um Lo a d in g
Pr e miu mLin e
Pre mi u mIn v o ic eRe l at e dA ct io n
Pr em iu mIn v o ic e Lin e
Pre mi u mIn v o ic e
Pr e miu mDi sc o u n tOr Lo a d in g
Pre miu mDi sc o u n t
Pre miu mCa n c el la ti on Li ne
Pr emi u mAn d Com mis si o n Re v e rs ib l ePr emi u mLin e
Pr e miu m
Po lic y Ho ld e r
Pol ic y
Ph y s ic al Per so n
Ph y s ic a lI nj u ry
Per s on a lLi ab i lit y In s u ra n ce Co n tr ac t
Per ma n en t Dis a bi li ty Co v er
Pe rma n e n tDi sa b il ity
Pa rt y Rol e
Pa rt y
Pa rt ia lTe mp o ra ry Di sa b il ity
Par it yCo mmi tt e e
Own e r
Oth e r Pr iv a te Bu il di n gOth e rPo li cy
Ot h er In s u ra n c eRe d u ct io n
O pp o n e n tTh ir d Pa r ty
Old Pr emi u m
Off ic ia lAd d r es s
O ffi c eTe le p h o ne
Of fe n ce Pe n di n g
Not if ie r
No n Reg i st e re d Le g a lPe rs o n
N on M o to ri se d Ro a dVe h ic le
Net Pre miu mN on i nd e x ed
Ne tPre mi umI n d ex e d
Net Pre miu m
NCBYe a rM in u s3
N CBYe a rM in u s 2
NCBRef er e n ce
NCBPre v io u s N CBLe v e l
NCBAt Un d er wr iti n g
NCBAt Pr e v io u sI n su r er
NCBAc tu a l
Mu t u al Hea l th In s u re r
Mo to r Thi rd Pa rt y Li a bi lit y
M o to r The f t
Mo to r Par ti al Co mp re h en s iv e
M o to r is ed Ro a d Veh i cl e
Mo t o rI ns u ra n c eCo n t ra c t
M ot o rFu ll Com pr e he n s iv e
Mo t o rFir e
M ot o rCy c le Ce rt ifi c at e
Mo t or Co mp re h e ns iv e
Mo r tg a g eLo a n Fi le
Mo ra lD ama g e
Mo b il eTe le p h o n e
M is c el la n eo u s Tra n s ac ti o n Li n e
Me d ic al Do ct o r
M at er ia lD ama g e
Ma n a g eme n t Act
Ma il Sec o n d ar y Ad d re s s
M ai lPri ma ry Ad d re s s
M a ch i ne r y /Ap p li an c e s
Lo ad Ca p a ci ty
Lif ePr o fit Sh ar in g
Lif eI n su r a nc e Co n tra c t
Le ga l Pe r so n
Leg a lEx p en s e sI n su r a nc e Co n tra c t
Lea s in g Co mp an y
Las t Pr e miu m
La s tCo mmi ss i on
La n d
Kfk Re q ue s tFo r Ref u n d
Kf kOp e ra t io n
Kf kDe b it OrCr e di tSt at e men t
J o bCa t eg o r ySe n io r ity
I n te rv e n in g Au th o r it y
In te r me di ar y
In s u re r
In s u re d Va lu eBa s e dD ed u c ti bl e
I n su r ed Pe rs o n
I ns u r ed
In s u ra n ce Pro v id e r
I nd i vi d u al In s u ra n ce Co n tr ac t
I nd e x
In d e mn ifi c at io n
I d en t it y Ca r d
Hu n ti ng Li ce n c e
Hun t in g
Ho s p it al
Ho me Te le p ho n e
Ho l id a yPa y Bu re a u
Hea d Of fic e Ad d re s s
Ha ul a ge Tr ai le r
Gua r an t ee
Gr o u pI n s ur a nc e Co n tr ac t
Gr o u p
Gro s s Pre miu m
Gr ee n Ca rd
Ge ne r al Life
Ge n e ra lFi re
Ge n e ra lCo n d it io n sW o rd in g
Gen e ra l Co n d it io n
Ge ne r al Civ il Lia bi li ty
Gar a ge
Fu tu r eA dd r es s
Fle e t
Firs t Pr e miu m
Firs t Comm is si o n
Fir s tAi d Su p p li er
Fir eI n su r an c e Co nt ra c t
Fi na n c ia lTr an s a ct io n Lin e
Fi n an c ia lDa ma g e
Fin a n c eCo mp a n y
Fil e Ad min is t ra to r
Fid e li ty Bo n us
Fe e
Fax
Fa mily A llo wa n c eBu r ea u
Fa ct u al Gro u p
Ex te r n al Rol e
Ex te n si o n
Ex p e ns e
Ev e nt Ad d re s s
Eu r o pe a n Ac ci d en t No ti fic a ti o nFo r m
En g in e
En d In d e x
Empl o y ee Be n ef it sCo n tr a ct
e M ai l
D wel lin g
Dr iv i ng Lic e n ce W it hd r a wal
Dr iv in g Lic e nc e
Dr iv e r
Do c u men t Exc h an g e De ta il s
Do c u me nt
Di sa b il it y
D ire c tPa y me nt
Ded u c ti bl e
D ec la r at io n
Dec e a se Pro f it Sha r in g
Dat a Ent ry Op e ra t or
Da ma ge d Ob je c t
D ama g e Amo u n tBa se d De d u ct ib le
Da ma g e
Da ily Al lo wa n ce
Cur re n tA cc o u n tLin e
Cu rr en t Ac co u n t
Cr ed i to r
Co v er No te
Co v er Limit
Co v e r
Co up l e
Co st /Ch a rg e
Co Po li cy Ho ld e r
Co n v ic t io n
Co n tr ac t Do cu me n t
Co nt ra c tCh a n g e
Co n t ra ct
Co nt e n ts
Co nd i tio n
Co mp a n ySe n io r ity
Co mmu n ic a tio n Ch a n ne l
Com mo nLa w
Co mmi ss io n Lin e
Com mis si o n
Co mmer c ia lM ot o ri se d Ro ad Ve h ic le
Co mme r ci al Los s
Com mer c ia lI ns u ra n c e Co n t ra c t
Co mmer c ia lDi sc o u n t
Com mer c ia lBu il di n g
Co ll e ct io n
Co in s u re r
Cla u s e
Cla imLin e
Cla ima n t
Cla im
Civ ilLi ab i lit y Ho rs e sA nd Ca rr ia g e s
Cir cu ms ta n c eCo l u mn
Cir c um st an c e
Cer ti fy in g M ed i ca lDo c t or
Ca ra v a n
Ca p it al
Ca n ce ll at io n
Ca lc u la ti o n
Bu r g la ry /Th e ft
Bui ld in g
Bra n ch Se n io ri ty
Br a n ch Of fi ce Ad d re s s
Bor d er e au To t al Lin e
Bo rd e re a u Li n e
Bo r d er ea u
Bo n u s Rec e p ti on De c la ra t io n
Bo n u s
Bill in g Ad d re s s
Bi d/ Off e r
Bi cy c le
Ben e fi c ia ry
Bas e Pre miu m
A ss e s so r
An n ul me nt
An n u al Pre miu mNo n I nd e x ed
An n ua l Pr e miu mIn d e x edAn n u al Pre miu m
A n nu a lI n d emn i fic a ti o n
Ame n d men t
Air Ve hi c le
Ad va n t ag e In Ki nd Re c ep t io n De cl a ra ti on
Ad v an t ag e In Ki n d
Ad d re ss
Ad d On In s u re r
Ad d it io n a lOc c up a t io n
Ad d it io n al Ac ti vi ty
Ac tu a lI ns u r er Of
Ac tu a lI n de x
Act iv i ty Exe c u tio n Ad d r es s
Ac ti on
Ac qu i si ti on Co mmi ss i on
Ac co u n ti n gDo c u me n tLin e
Ac c o u nt in g Do c u men t
Ac c id e n ts Sin c eA dj u st me n t
Ov er v ie w/Co n c ep t u al
Top-8
Top-7 after first collapsing isA hierarchies
1-1
1-1
role"c"
1-1
1-1
premium"c"
1-11-1 PremGuarantee"c"
1-1
1-1
partyRole"c"
1-1 1-1participants
1-1
1-1
located"c"
1-1
1-1insurObj"c"
1-1
1-1
hasAddress"c"
1-1
1-1
contractObj"c"
1-1
1-1
collapseInsuranceObject11
1-1
1-1
address"c"
RelatedParty"c"
Premium"c"
PartyRole"c"
InsuranceObject"c"
Guarantee"c"
Contract"c"
Address1
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 88
Formal Evaluation Methods(TREC/INEX-like)
Steps:
1. Derive the Ideal Ranking of the elements of a schema
– by aggregating the top-k lists provided by several experts on that schema
2. Compare an automatically derived ranking w.r.t. the ideal ranking
based on appropriate metrics
– ... metrics for evaluating the effectiveness of Information Retrieval Systems
How to visualise these top-k diagrams ?
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 90
How to visualise these top-k diagrams ?
• For drawing automatically the top-k diagrams we combine the spring-model [Eades84] with the magnetic-spring model
[SugiyamaMisue94] in a way that is appropriate for ER diagrams.
• We view an ER diagram as a mechanical system
– Force model A
– Force model B
• Drawing algorithm
– Algorithm that simulates the mechanical system
– Its seeks for a configuration with locally minimal energy
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 91
Force Model A
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 92
Force Model A: Computing the exerted forces
• Force exerted on an entity type:
),(),(),()()(,)(
ij
ieCONISAjeij
jeieEjeij
ieCONjei eeheegeefeF ∑∑∑
∈≠∈∈
++=
String force
from connections
Electrical repulsion
from every other particle
Magnetic (rotational)
force from ISA connections
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 93
Force Model A: Configuration parameters
∑∈
−−=
)(,,
),()),(()(
ieCONje ji
ijjiji
sjiix
ppd
xxLppdKef
∑≠∈
−=
iejeEje ji
ji
ji
eji
ixppd
xx
ppd
Keg
,2
,
),(),()(
∑∑∈∈
−+
−=
)( ,)( ,
)(
ieCONNSUBje ji
ijm
ieCONNSUPje ji
ijmix
L
xxK
L
xxKeh
),(),(),()()(,)(
ij
ieCONISAjeij
jeieEjeij
ieCONjei eeheegeefeF ∑∑∑
∈≠∈∈
++=
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 94
The role of the forces f, g, and hand of the configuration parameters
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 95
Force Model B
• Relationship types are also considered as particles in order to discourage overlaps.
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 96
B: Drawing ER diagramsExperimental Evaluation> Multiple Isa Hierarchies
1-10-N
R_3
1-10-NR_2
1-1
0-NR_1
1-10-N
R
ENTITY_9
ENTITY_8
ENTITY_7ENTITY_6
ENTITY_5
ENTITY_4
ENTITY_3
ENTITY_2
ENTITY_11
ENTITY_10
ENTITY_1
ENTITY
1-1
0-N
R_9
1-1
0-NR_8
1-1
0-N
R_7
1-1
0-N
R_6
1-1
0-N
R_5
1-10-N R_4
1-1
0-N
R_3
1-1
0-N
R_2
1-1
0-N
R_10
1-1
0-N
R
ENTITY_9
ENTITY_8
ENTITY_7
ENTITY_6
ENTITY_5
ENTITY_4
ENTITY_3
ENTITY_2
ENTITY_1
ENTITY
1-1
0-N
R_9
1-1 0-N
R_8
1-1
0-N
R_7
1-1
0-N
R_6
1-1
0-N
R_5
1-10-N R_4
1-1
0-N
R_3
1-1
0-NR_2
1-1
0-N
R_10
1-1
0-N
R
ENTITY_9
ENTITY_8 ENTITY_7
ENTITY_6
ENTITY_5ENTITY_4
ENTITY_3ENTITY_2
ENTITY_1
ENTITY
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 97
Experimental EvaluationFM A vs FM B
1-1
0-N
R_9
1-1
0-N
R_8
1-1
0-N
R_7
1-1
0-N
R_6
1-1
0-N
R_5
1-1
0-N
R_4
1-10-NR_31
1-1
0-N
R_30
1-1
0-N
R_3
1-1
0-N
R_29
1-10-N R_28
1-1
0-N
R_27
1-1
0-N
R_26
1-1
0-N
R_251-1
0-N
R_24
1-10-N R_23
1-1
0-N
R_22
1-1
0-N
R_21
1-1 0-NR_20
1-1
0-N
R_2
1-1
0-N
R_19
1-1
0-N
R_18
1-1
0-N
R_17 1-1
0-N
R_16
1-1
0-N
R_15
1-1 0-NR_14
1-1
0-N
R_13
1-1 0-NR_12
1-1
0-N
R_11
1-1
0-N
R_10
1-1
0-N
R_1
1-1
0-N
R
ENTITY_9
ENTITY_8
ENTITY_7
ENTITY_6
ENTITY_5
ENTITY_4
ENTITY_3
ENTITY_2
ENTITY_16
ENTITY_15
ENTITY_14ENTITY_13
ENTITY_12
ENTITY_11
ENTITY_10
ENTITY_1
ENTITY
1-1
0-N
R_9
1-1
0-N
R_8
1-1
0-N
R_7
1-1
0-N
R_6
1-1
0-N
R_5
1-1
0-N
R_4
1-1
0-N
R_31
1-1
0-N
R_30
1-1
0-N
R_3
1-1
0-N
R_29
1-1
0-N
R_28
1-1
0-N
R_27
1-1
0-N
R_26
1-1
0-N
R_25
1-1
0-N
R_24
1-1
0-N
R_23
1-1
0-N
R_22
1-1
0-N
R_21
1-1
0-N
R_20
1-1
0-N
R_2
1-1
0-N
R_19
1-1
0-N
R_18
1-1
0-N
R_17
1-1
0-N
R_16
1-1
0-N
R_15
1-10-NR_14
1-1
0-N
R_13
1-10-NR_12
1-1
0-N
R_11
1-1
0-N
R_10
1-1
0-N
R_1
1-1
0-N
R
ENTITY_9
ENTITY_8
ENTITY_7
ENTITY_6
ENTITY_5
ENTITY_4
ENTITY_3
ENTITY_2
ENTITY_16
ENTITY_15
ENTITY_14
ENTITY_13
ENTITY_12
ENTITY_11
ENTITY_10
ENTITY_1
ENTITY
• FM B
– The tentangles of binary relationship types are unnecessarily not aligned.
– Computationally more expensive
• FM Α
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 98
Experimental Evaluation> On bigger diagrams
0-N
0-1
ιmis
0-N
1-1ΰ
0-N
1-1
vers
0-N0-1Type
1-N
0-N
titulaire
0-N
0-N
supplι ant
0-N
1-1
secrι taire
0-1
1-2
rι dige
0-N
0-1
reηu
requiert
0-N
est requis
0-1
requiert
0-N
1-1
prι sident
0-N
1-1
Propose
0-N
1-1
pour
0-N 0-N
participe
1-1
0-N
organise
1-10-N
organise
0-N
0-N
offre rι elle
0-N
0-N
offre
0-N
1-1
Nationalitι
0-N
1-1
naissance
0-N
0-N
membre
0-N
1-1
inscrit
0-3
0-N
inscription-examen
0-N
0-1
etude secondaire
1-1
0-N
de
0-N
1-1de
0-N
1-1
de
0-N
1-1de
0-1
0-N
de
0-N
1-1de0-N
1-N
de
0-N
1-1
de 0-N
1-1
de
0-N
1-1
de
0-N
1-1
de
0-N
1-1
de
0-N
1-1
de
0-N
1-1
de
0-N1-1
de
0-N
1-1
de
0-N
0-1
de
0-366
1-1
de
0-N
1-1
de
1-1
1-1
de
0-N
0-1
de
0-N
1-1
de
0-N
1-1
de
0-N
1-1
de
0-N
0-1
derniere inscription univ
1-1
0-N
de
0-N
0-1
dans
0-N
0-1
dans
0-3
1-Nconcerne
0-N
0-1compose
0-N
0-N
assistant-TP
0-N
0-N assistant-cours
1-1
1-2
adresse
P
P
T
P
UNIVERSITE
UNITE
TYPE_HUMANITE
SESSIONROLE
PROG_ECHANGE
PAYS
MOUVEMENT_COMPTABLE
MEMOIRE
LOCALITE
INSCRIPTION-COURS
FUNDP
FRAIS D'INSCRIPTION
FACULTE
EXTERNE
ETUDIANT
ETUDE_SUPERIEURE
ETUDE UNIVERSITAIRE
ETABLISSEMENT
ENSEIGNANT
DOCUMENT RECU
DOCUMENT EMIS
DOCUMENT
DEPARTEMENT
DELIBERATION
COURS-TYPE
COURS-EXCEPT
COURS
COMPTE_ETUDIANT
COMPTE
COMMUNE
CLASSE-TYPE
CLASSE
CALENDRIER
AUTRE_COMPTE
ANNEE_ACADEMIQUE
ADRESSE
ADMINISTRATIF
ACTIVITE POST SECONDAIRE
1-1
0-N
R_99
1-1
0-N
R_98
1-1
0-N
R_97
1-1
0-N
R_96
1-1
0-N
R_95
1-1
0-N
R_94 1-1
0-N
R_93
1-1
0-N
R_92
1-1
0-N
R_91
1-1
0-N
R_90
1-1
0-N
R_9
1-1
0-N
R_891-1
0-N
R_88
1-1
0-N
R_87
1-1
0-N
R_86
1-1
0-N
R_85
1-1
0-N
R_84 1-1
0-N
R_83
1-1
0-N
R_82
1-1
0-N
R_81
1-1
0-N
R_80
1-1
0-N
R_8
1-1
0-N
R_79
1-1
0-N
R_78
1-1
0-N
R_77
1-1
0-N
R_76
1-1
0-N
R_75
1-1
0-N
R_74
1-1
0-N
R_73
1-1
0-N
R_72
1-1
0-N
R_71
1-1
0-N
R_70
1-1
0-N
R_7
1-1
0-N R_69
1-1
0-N
R_68
1-1
0-N
R_67
1-1
0-N
R_66
1-1
0-N
R_65
1-1
0-N
R_64
1-1
0-N
R_63
1-1
0-N
R_62
1-1
0-N
R_61
1-1
0-N
R_60
1-1
0-N
R_6
1-1
0-N
R_59
1-1
0-N
R_581-1
0-N
R_57
1-1 0-N
R_56
1-1
0-N
R_55
1-1
0-N
R_54
1-1
0-N
R_53
1-1
0-N
R_52
1-1
0-N
R_51
1-1
0-N
R_50
1-1
0-NR_5
1-1
0-N
R_49
1-1
0-N
R_48
1-1
0-N
R_47
1-1
0-N
R_46
1-1
0-N
R_45
1-1
0-N
R_44
1-1
0-N
R_43
1-1
0-N
R_42
1-1
0-N
R_41
1-1
0-N
R_40
1-1
0-NR_4
1-1
0-N R_39
1-1
0-N
R_38
1-1
0-N
R_37
1-1
0-N
R_36
1-1
0-N
R_35
1-1
0-N
R_34
1-10-N
R_33
1-1
0-N
R_32
1-1
0-N
R_31
1-1
0-N
R_30
1-1
0-N
R_3
1-1
0-N
R_29
1-1
0-N
R_28
1-1
0-N
R_27
1-1 0-N
R_26
1-1
0-N
R_25
1-1
0-N
R_24
1-1
0-N
R_23
1-1
0-N
R_22
1-1
0-N
R_21
1-1
0-N
R_20
1-1
0-N
R_2
1-1
0-N
R_19
1-10-N
R_18
1-1
0-N
R_17
1-1
0-N R_16
1-1
0-N
R_15
1-1
0-N
R_14
1-1
0-N
R_13
1-1
0-NR_12
1-1
0-N
R_111
1-1
0-N
R_110
1-1
0-N
R_11
1-1
0-N
R_109
1-1
0-N R_108
1-1
0-N
R_107
1-10-N
R_106
1-1
0-N
R_105
1-1
0-N
R_104
1-1
0-NR_103
1-1
0-N
R_102
1-1
0-N
R_101
1-1
0-N
R_100
1-1
0-N
R_10
1-1
0-N
R_1
1-1
0-N
R
ENTITY_9
ENTITY_8
ENTITY_7
ENTITY_6
ENTITY_5
ENTITY_47
ENTITY_46
ENTITY_45
ENTITY_44
ENTITY_43
ENTITY_42
ENTITY_41
ENTITY_40
ENTITY_4
ENTITY_39
ENTITY_38
ENTITY_37
ENTITY_36
ENTITY_35
ENTITY_34
ENTITY_33
ENTITY_32
ENTITY_31
ENTITY_30
ENTITY_3
ENTITY_29 ENTITY_28
ENTITY_27
ENTITY_26
ENTITY_25
ENTITY_24
ENTITY_23
ENTITY_22
ENTITY_21
ENTITY_20
ENTITY_2
ENTITY_19
ENTITY_18
ENTITY_17
ENTITY_16
ENTITY_15
ENTITY_14ENTITY_13
ENTITY_12
ENTITY_11
ENTITY_10
ENTITY_1ENTITY
Artificial diagram Real ER diagram
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 99
Experimental Evaluation> On real diagrams
Top-5: FM A (after 2 different initial placements)
0-N
0-N
0-Nsupplies
0-N
0-N
Stock Movement
replaces
0-1
replacedBy
0-1replacement
0-N
0-N
POSITIONS/ACTIVITIES/MONTHS
0-N
0-N
0-N
path Calculation
0-N
1-N
Operator
0-N
1-1description
0-1
describes
0-NNotice
0-N
0-N
Max Weight/Channel
0-N
0-N
0-N
follow-up activity
Bonus
0-N
0-N
0-N
EARLY BIRD
0-N
0-N
DefaultCost
partOf
0-N
0-Ndecomposition
origin
0-N
destination
0-NCustoms costs
0-N
0-N
country language
0-N
1-1
cn/pay
0-N
0-N
0-N
channelCalculation
0-N
0-N
accumulated movements
State of Progress
SectorOfActivity
Item
Country
Channel
0-N
0-N
0-N
supplies
0-N
0-N
Stock Movement
replaces
0-1
replacedBy
0-1replacement
0-N
0-N
POSITIONS/ACTIVITIES/MONTHS
0-N
0-N
0-N
path Calculation
0-N
1-N
Operator 0-N1-1description
0-1
describes
0-NNotice
0-N
0-N
Max Weight/Channel0-N
0-N
0-N
follow-up activity
Bonus
0-N
0-N
0-N
EARLY BIRD
0-N
0-N
DefaultCost
partOf
0-N
0-Ndecomposition
origin
0-N
destination
0-NCustoms costs
0-N
0-N
country language
0-N
1-1
cn/pay 0-N
0-N
0-N
channelCalculation
0-N
0-N
accumulated movements
State of Progress
SectorOfActivity
ItemCountry
Channel
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 100
Experimental Evaluation> On real diagrams
Top-5: FM B
0-N
0-N
0-N
supplies
0-N
0-N
Stock Movement
replaces
0-1
replacedBy
0-1
replacement0-N
0-N
POSITIONS/ACTIVITIES/MONTHS
0-N
0-N
0-N
path Calculation
0-N
1-N
Operator
0-N
1-1
description
0-1
describes
0-NNotice
0-N
0-N
Max Weight/Channel
0-N
0-N
0-N
follow-up activity
Bonus
0-N
0-N
0-N
EARLY BIRD
0-N
0-N
DefaultCost
partOf
0-N0-N
decomposition
origin
0-N
destination
0-N
Customs costs
0-N
0-N
country language
0-N
1-1
cn/pay
0-N
0-N
0-N
channelCalculation
0-N
0-N
accumulated movements
State of Progress
SectorOfActivity
Item
Country
Channel
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 101
Experimental Evaluation> On real diagrams
Top-11: FM B (dense diagram)
Solutions:
• Play with the configuration parameters
– local density-based configuration of parameters (future research)
• or just scale up the entire diagram
0-N
0-N
SUIVRE LE DOSSIER
1-N
0-N
SUIVI OPERATRICE
0-N
0-N
SUIVI LIVRAISON COLIS
0-N
0-N
SUIVI DES TOPS CLIENT 0-N
0-N
1-1
SUIVI COMMANDES
0-N
0-N
0-NSUIVI ACTIVITE
1-N
1-N
STAT FOURNISSEUR MENSUELLES
0-N
0-N
simulation prix de vente
0-N
0-N
rιexpιdition colis
est remplace
0-1
remplace
0-1Remplacement
date dιbut de validitι
0-N
date de fin de validitι
0-N
0-N remise fin annιe fournisseur
0-N
0-N
POSITIONS/ACTIVITE/MOIS
0-N
0-N
POS-DTE
0-N
1-N
pondιration
0-N
0-N
Poids max/canal
0-N
1-N
PAYS LANGUE
0-N
0-N
0-N
PAYS ENSEIGNE LANGUE
0-N
1-1
pays du fournisseur
0-N
1-1
pays de la devise0-1
0-N
Pays d'origine dιcrire
0-N
description
0-1
1-1
0-N
0-N Notice
0-N
0-N
Mouvement de stock
0-N
0-NDate de depot
1-1LOT
1-1
0-N
LIVRAISON COLIS
1-N
1-1
HISTORIQUE COLIS
0-N
0-NFRAIS PAR DEFAUT
pays destination
0-N
pays d'origine
0-NFRAIS DE DOUANE
1-1
0-N
fournisseur/pays
1-1
0-N
fournisseur/langue
0-N
0-N
FORMULE PAR PAYS/LANGUE
0-N
0-N
0-NCadeau associι
0-N
EARLY BIRD
article vpc piθce
0-N
article vpc ensemble
0-Ndιcomposition d'un ensemble
0-N
0-N
dte/pay1-1
0-N
Document par langue
0-N
0-N
cumul mouvement systeme mesure
0-N
0-N
Composition
0-N
0-N
colcom/suicol
0-N
1-1
col/eta
1-1
0-N
cn/suicol
1-1
0-N
cn/pay
1-1
0-N
cd/suicol
0-1
0-Ncd/eta1-1
0-N
Canal livraison
1-1
0-N
Canal commande
0-N
0-N
0-Ncalcul goulotte
0-N
0-N
0-N
calcul canal
1-10-Natvsec/cd
0-N
0-1
artvpc/frn
0-N
1-N
artref/tva
0-N
1-N
artref/douref
1-N
0-N
artref/atvsec
0-N
1-1
artref/artvpc
0-N
0-N
0-N
0-NALIMENTER
1-1
0-N
achfrn/lan
SECTEUR ACTIVITE
REFERENCE ARTICLE
PAYS
LANGUE
FOURNISSEUR
ETAT D'AVANCEMENT (LIGNE OU CO
DATE
COMMANDE
COLIS
CANAL
ARTICLE VPC
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 102
Experimental Evaluation> On real diagrams
Top-11: FM B (dense diagram)
0-N
0-N
SUIVRE LE DOSSIER
1-N
0-N
SUIVI OPERATRICE
0-N
0-N
SUIVI LIVRAISON COLIS
0-N
0-N
SUIVI DES TOPS CLIENT
0-N
0-N
1-1
SUIVI COMMANDES
0-N
0-N
0-N
SUIVI ACTIVITE
1-N
1-N
STAT FOURNISSEUR MENSUELLES0-N
0-N
simulation prix de vente
0-N
0-N
rιexpιdition colis
est remplace
0-1
remplace
0-1Remplacement
date dιbut de validitι
0-N
date de fin de validitι
0-N
0-N remise fin annιe fournisseur
0-N
0-N
POSITIONS/ACTIVITE/MOIS
0-N
0-N
POS-DTE
0-N
1-N
pondιration
0-N
0-N
Poids max/canal
0-N
1-N
PAYS LANGUE
0-N
0-N
0-N
PAYS ENSEIGNE LANGUE
0-N
1-1
pays du fournisseur
0-N
1-1
pays de la devise
0-1
0-N
Pays d'origine
dιcrire
0-N
description
0-1
1-1
0-N
0-N Notice
0-N
0-N
Mouvement de stock
0-N
0-NDate de depot
1-1
LOT
1-1
0-N
LIVRAISON COLIS
1-N
1-1
HISTORIQUE COLIS
0-N
0-N
FRAIS PAR DEFAUT
pays destination
0-N
pays d'origine
0-NFRAIS DE DOUANE
1-1
0-N
fournisseur/pays
1-1
0-N
fournisseur/langue
0-N
0-N
FORMULE PAR PAYS/LANGUE
0-N
0-N
0-N
Cadeau associι
0-N
EARLY BIRD
article vpc piθce
0-N
article vpc ensemble
0-Ndιcomposition d'un ensemble
0-N
0-N
dte/pay 1-1
0-N
Document par langue
0-N
0-N
cumul mouvement systeme mesure
0-N
0-N
Composition
0-N
0-N
colcom/suicol
0-N
1-1
col/eta
1-1
0-N
cn/suicol
1-1
0-N
cn/pay
1-1
0-N
cd/suicol
0-1
0-N
cd/eta
1-1
0-N
Canal livraison
1-1
0-N
Canal commande0-N
0-N
0-N
calcul goulotte
0-N
0-N
0-N
calcul canal
1-1
0-N
atvsec/cd
0-N
0-1
artvpc/frn
0-N
1-N
artref/tva
0-N
1-N
artref/douref
1-N
0-N
artref/atvsec
0-N
1-1
artref/artvpc
0-N
0-N
0-N
0-N
ALIMENTER
1-1
0-N
achfrn/lan
SECTEUR ACTIVITE
REFERENCE ARTICLE
PAYS
LANGUE
FOURNISSEUR
ETAT D'AVANCEMENT (LIGNE OU CO
DATE
COMMANDE
COLIS
CANAL
ARTICLE VPC
Less stiff springs Higher electrical repulsion
0-N
0-N
SUIVRE LE DOSSIER
1-N
0-N
SUIVI OPERATRICE
0-N
0-N
SUIVI LIVRAISON COLIS
0-N
0-N
SUIVI DES TOPS CLIENT
0-N
0-N
1-1
SUIVI COMMANDES
0-N
0-N
0-N
SUIVI ACTIVITE
1-N
1-N
STAT FOURNISSEUR MENSUELLES
0-N
0-N
simulation prix de vente
0-N
0-N
rιexpιdition colis
est remplace
0-1
remplace
0-1Remplacement
date dιbut de validitι
0-N
date de fin de validitι
0-N
0-Nremise fin annιe fournisseur
0-N
0-N
POSITIONS/ACTIVITE/MOIS
0-N
0-N
POS-DTE
0-N
1-N
pondιration
0-N
0-N
Poids max/canal 0-N
1-N
PAYS LANGUE
0-N
0-N
0-N
PAYS ENSEIGNE LANGUE
0-N 1-1pays du fournisseur0-N 1-1pays de la devise
0-1
0-N
Pays d'origine
dιcrire
0-N
description
0-1
1-1
0-N
0-N Notice
0-N
0-N
Mouvement de stock
0-N
0-N
Date de depot
1-1
LOT
1-1
0-N
LIVRAISON COLIS
1-N
1-1
HISTORIQUE COLIS
0-N
0-NFRAIS PAR DEFAUT
pays destination
0-N
pays d'origine
0-NFRAIS DE DOUANE 1-10-N fournisseur/pays
1-1
0-N
fournisseur/langue
0-N
0-N
FORMULE PAR PAYS/LANGUE
0-N
0-N
0-N
Cadeau associι
0-N
EARLY BIRD
article vpc piθce
0-N
article vpc ensemble
0-Ndιcomposition d'un ensemble
0-N
0-N
dte/pay
1-10-N
Document par langue
0-N
0-N
cumul mouvement systeme mesure
0-N
0-N
Composition
0-N
0-N
colcom/suicol
0-N
1-1
col/eta
1-1
0-N
cn/suicol
1-1
0-N
cn/pay
1-1
0-Ncd/suicol
0-1
0-N
cd/eta
1-1
0-N
Canal livraison
1-1
0-N
Canal commande
0-N
0-N0-N
calcul goulotte
0-N0-N
0-N
calcul canal1-1
0-N
atvsec/cd 0-N
0-1
artvpc/frn
0-N
1-N
artref/tva
0-N
1-N
artref/douref
1-N
0-N
artref/atvsec
0-N
1-1
artref/artvpc
0-N
0-N
0-N
0-N
ALIMENTER
1-1
0-N
achfrn/lan
SECTEUR ACTIVITE
REFERENCE ARTICLE
PAYS
LANGUE
FOURNISSEUR
ETAT D'AVANCEMENT (LIGNE OU CO
DATE
COMMANDE
COLIS
CANAL
ARTICLE VPC
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 103
Experimental Evaluation> On real diagrams
Top-11: After manual rectification/beautification
0-N0-N
SUIVRE LE DOSSIER
1-N
0-N
SUIVI OPERATRICE
0-N
0-N
SUIVI LIVRAISON COLIS
0-N0-N
SUIVI DES TOPS CLIENT
0-N0-N
1-1
SUIVI COMMANDES
0-N
0-N 0-N
SUIVI ACTIVITE
1-N
1-N
STAT FOURNISSEUR MENSUELLES
0-N
0-N
simulation prix de vente
0-N
0-N
rιexpιdition colis
est remplace
0-1
remplace
0-1
Remplacement
date dιbut de validitι
0-N
date de fin de validitι
0-N
0-N
remise fin annιe fournisseur
0-N 0-N
POSITIONS/ACTIVITE/MOIS
0-N
0-N
POS-DTE
0-N
1-N
pondιration
0-N
0-N
Poids max/canal
0-N1-N
PAYS LANGUE
0-N0-N
0-N
PAYS ENSEIGNE LANGUE
0-N
1-1
pays du fournisseur
0-N
1-1
pays de la devise
0-1
0-N
Pays d'origine
dιcrire
0-N
description
0-1
1-1
0-N
0-N
Notice
0-N0-N
Mouvement de stock
0-N0-N
Date de depot
1-1
LOT
1-10-N
LIVRAISON COLIS
1-N
1-1
HISTORIQUE COLIS
0-N
0-N
FRAIS PAR DEFAUT
pays destination
0-Npays d'origine
0-N
FRAIS DE DOUANE
1-1
0-N
fournisseur/pays1-1
0-N
fournisseur/langue
0-N0-N
FORMULE PAR PAYS/LANGUE
0-N0-N
0-N
Cadeau associι
0-N
EARLY BIRD
article vpc piθce
0-N
article vpc ensemble
0-N
dιcomposition d'un ensemble
0-N
0-N
dte/pay
1-1
0-N
Document par langue
0-N
0-N
cumul mouvement systeme mesure
0-N0-N
Composition
0-N
0-N
colcom/suicol
0-N
1-1
col/eta
1-10-N
cn/suicol
1-10-Ncn/pay
1-1
0-N
cd/suicol
0-1 0-Ncd/eta
1-1
0-N
Canal livraison
1-1
0-N
Canal commande
0-N
0-N
0-N
calcul goulotte
0-N
0-N0-N
calcul canal
1-1
0-N
atvsec/cd
0-N
0-1
artvpc/frn
0-N
1-N
artref/tva
0-N
1-N
artref/douref
1-N0-N artref/atvsec 0-N 1-1artref/artvpc
0-N
0-N
0-N
0-N
ALIMENTER
1-1
0-N
achfrn/lanSECTEUR ACTIVITE
REFERENCE ARTICLE
PAYS
LANGUE
FOURNISSEUR
ETAT D'AVANCEMENT (LIGNE OU CO
DATE
COMMANDE
COLIS
CANAL
ARTICLE VPC
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 104
Extending the CASE tool DB-MAINwith a semiautomatic conceptual drawing approach
select
move nail unnail collapse draw Drawing algorithms
Semantic-based
selection algorithms
manual
...
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 105
Concluding Remarks
0-N0-N
SUIVRE LE DOSSIER
1-N
0-N
SUIVI OPERATRICE
0-N
0-N
SUIVI LIVRAISON COLIS
0-N0-N
SUIVI DES TOPS CLIENT
0-N0-N
1-1
SUIVI COMMANDES
0-N
0-N 0-N
SUIVI ACTIVITE
1-N
1-N
STAT FOURNISSEUR MENSUELLES
0-N
0-N
simulation prix de vente
0-N
0-N
rιexpιdition colis
est remplace
0-1
remplace
0-1
Remplacement
date dιbut de validitι
0-N
date de fin de validitι
0-N
0-N
remise fin annιe fournisseur
0-N 0-N
POSITIONS/ACTIVITE/MOIS
0-N
0-N
POS-DTE
0-N
1-N
pondιration
0-N
0-N
Poids max/canal
0-N1-N
PAYS LANGUE
0-N0-N
0-N
PAYS ENSEIGNE LANGUE
0-N
1-1
pays du fournisseur
0-N
1-1
pays de la devise
0-1
0-N
Pays d'origine
dιcrire
0-N
description
0-1
1-1
0-N
0-N
Notice
0-N0-N
Mouvement de stock
0-N0-N
Date de depot
1-1
LOT
1-10-N
LIVRAISON COLIS
1-N
1-1
HISTORIQUE COLIS
0-N
0-N
FRAIS PAR DEFAUT
pays destination
0-Npays d'origine
0-N
FRAIS DE DOUANE
1-1
0-N
fournisseur/pays1-1
0-N
fournisseur/langue
0-N0-N
FORMULE PAR PAYS/LANGUE
0-N0-N
0-N
Cadeau associι
0-N
EARLY BIRD
article vpc piθce
0-N
article vpc ensemble
0-N
dιcomposition d'un ensemble
0-N
0-N
dte/pay
1-1
0-N
Document par langue
0-N
0-N
cumul mouvement systeme mesure
0-N0-N
Composition
0-N
0-N
colcom/suicol
0-N
1-1
col/eta
1-10-N
cn/suicol
1-10-Ncn/pay
1-1
0-N
cd/suicol
0-1 0-Ncd/eta
1-1
0-N
Canal livraison
1-1
0-N
Canal commande
0-N
0-N
0-N
calcul goulotte
0-N
0-N0-N
calcul canal
1-1
0-N
atvsec/cd
0-N
0-1
artvpc/frn
0-N
1-N
artref/tva
0-N
1-N
artref/douref
1-N0-N artref/atvsec 0-N 1-1artref/artvpc
0-N
0-N
0-N
0-N
ALIMENTER
1-1
0-N
achfrn/lanSECTEUR ACTIVITE
REFERENCE ARTICLE
PAYS
LANGUE
FOURNISSEUR
ETAT D'AVANCEMENT (LIGNE OU CO
DATE
COMMANDE
COLIS
CANAL
ARTICLE VPC
10 minutes
U. of Crete, Information Systems Analysis and Design Yannis Tzitzikas, Fall 2005 106
How to Tame a Very Large ER Diagram (using Link Analysis and Force-Directed Placement Algorithms)
Concluding Remarks• The diagrams of large schemas are very difficult to understand and visualize
• Top-k diagrams that are derived by (B)EntityRank can aid the understanding, the visualization, and the drawing of such diagrams.
• The experiments on real diagrams were successful.
• We can draw top-k diagrams successfully (for small k)
Further research
• Ranking
– BEntityRank and User Feedback
– Test Collections for Evaluations
– Take into account multiplicity constraints?
• Drawing
– Local-Density Adaptation (EntityRank scores might help)
Recommended