INSIDE! $7.00 U.S. SPECTRUM INTERNA TI ONAL · Is $7.00 U.S. ® INTERNA TI ONAL THE BUSINESS COMPUTER MAGAZINE JANUARY/FEBRUARY 2005 • AN IDBMA, INC. PUBLICATION SPECTRUM INSIDE!

Is

$7.00 U.S.

®

I N T E R N A T I O N A L

JANUARY/FEBRUARY 2005 • AN IDBMA, INC. PUBLICATIONTHE BUSINESS COMPUTER MAGAZINE

SPECTRUMSPECTRUMI N S I D E ! Blast From the Past: Getting Current With People Who Graced Spectrum’s Cover

MultiValue

DinosaurWhen It Comes to Information

Technology Controls and Audit Procedures

aMultiValue


Technology Controls andAudit Procedures?

COBIT: A Guide to Compliance With New Government Regulations


INTERNATIONAL SPECTRUM JANUARY/FEBRUARY 2005 ◆ 3

SPECTRUM®J A N U A R Y / F E B R U A R Y 2 0 0 5


6 From the Inside

32 Newsmakers

32 New Products

INTERNATIONAL SPECTRUM is

a registered trademark and

MultiValue is a trademark of

IDBMA Inc. All other registered

trademarks and trademarks are

the property of the respective

trademark holders.

DE

PA

RT

ME

NT

S

16 WHERE ARE THEY NOW? Two Decades of Spectrum Magazine: Influential People That Made the Cover Take a walk down memory

lane with this peek into the past and current lives of 17 people who have appeared onthe cover of Spectrum. BY LEANNE GREEN

24SERVICE ORIENTED BUSINESS APPLICATIONS: DISSIMILAR APPLICATIONS WORKING TOGETHER Service Oriented Business Application

(SOBA) is actually more than another acronym and building Web-based applications usingSOA standards. If you want to take advantage of Web Services already in existence, removebarriers to doing business with customers and suppliers, or attract the next crop of develop-ers, then you are probably interested in learning more about SOBA. BY LEE BURSTEIN

41USING BASIC AND PERL TO SEND ELECTRONIC MAIL MESSAGESA common problem encountered with legacy systems is they aren’t designed

to generate and send status messages regarding reports. With a little ingenuity, youcan accomplish a better flow of information by utilizing either Basic or Perl to sendelectronic messages. BY ROBERT COLTUN

14 WEBONOMICS 101: NEWS THAT GRABS YOU — EXPLAINING RSSA few years back, Netscape introduced Rich Site Summary (RSS), and the

concept has moved forward since then. Find out how companies can use RSS for theirbenefit. BY MELVIN SORIANO

34 REVELATION TECH TIPS: OPENINSIGHT FOR LINUX, AN OVERVIEWAs the Linux desktop environment grows more usable, Revelation Software

capitalizes on new opportunities in the Linux world. BY MIKE RUANE

26 UNIVERSE DYNAMICALLY HASHED FILES: TUNING PARAMETERS, PART 6The concept that dynamic files are maintenance free is a misconception held

by many database administrators and prompts the question, “If dynamic files don’t need anymaintenance, why are there so many tuning parameters?” This article discusses the tuningparameters for dynamically hashed UniVerse files. BY JEFF FITZGERALD AND PEGGY LONG

40 SB+ 101: SB+ BASICS: TOOLS, FILE CREATION AND DELETIONIn the seventh article in this series, the author covers SB+ tools, file creation

and deletion. BY DANNY PASSIG

6 WHERE IN THE WORLD IS VMARK? Unbeknownst to many in the MultiValuecommunity, VMark is alive and prospering in many parts of the world. Find out

what’s been going on behind the scenes with the company that many thought had metits demise when it merged with Unidata and became Ardent.

46 MVTOOLCHEST: MVTOOLBOX MAKES PROGRAMMERS’ TASKS EASY AND EFFICIENT The basic programming utilities supplied with

MultiValue databases have not been significantly improved for over three decades. Devel-oped by a programmer for programmers, mvToolbox changes all of that by providing com-prehensive programming tools that take the problems out of problem solving.

39 TECH TRENDS: RFID — WHAT IS IT? You’ve probably heard about RFIDin one form or another. Most of us, however, don’t really know what it can do,

how it works, or even what issues there may be when implementing it. BY NATHAN RECTOR

8 Taking the BiteOut of COBIT

Part 1: IT Controls: What AreThey and Who Wants Them?Have new federally mandatedIT controls and audit proceduresstemming from Sarbanes-Oxley legislation thrown yourMultiValue shop into a panic?How do you approach such a large task as implementingIT controls? This article is the first in a series that breaks down COBIT (Control Objectives for Informationand related Technology), acomprehensive frameworkthat addresses the entire setof business IT functions.

BY W. CLIFTON OLIVER

Is

$7.00 U.S.

®


JANUARY/FEBRUARY 2005 • AN IDBMA, INC. PUBLICATIONTHE BUSINESS COMPUTER MAGAZINE

SPECTRUMSPECTRUMI N S I D E ! Blast From the Past: Getting Current With People Who Graced Spectrum’s Cover

MultiValue



aMultiValue





4 ◆ INTERNATIONAL SPECTRUM JANUARY/FEBRUARY 2005

This is my favorite time of year when we're getting ready for the annual Spectrum Show. It's the time when we find out who's hot and who's not;what's new and what's not; who's doing well and who's crying the blues.

This year, the new third-party Web-based rapid application developmentsystems will be center stage. Conference attendees can get an overview ofthe capabilities of DesignBAIS, Visage, and WADE, three modern dayproducts competing for the hot MultiValue Web application development

market. Interested people can stay on for an extended technical sessionif after the overview they want to hear more in-depth detail.

The MultiValue to Microsoft Connection will also be more prominent in the show — thisyear's conference includes presentations on several fronts in many of these areas from Raining Data's PDP for .NET to how to interface your MultiValue data with Excel, Wordand Mail Merge.

In the category of “If you can't lick 'em; join ‘em,” there will be presentations on how totake your MultiValue system lock stock and barrel to Oracle or DB2, and still be able tocode in Data Basic or whatever you used to do before.

As the number of competing products in the MultiValue business intelligence tools marketincreases and the sophistication of drill-down models improves, your MultiValue enterprisecan now deliver truly astounding business intelligence to management. This year’s confer-ence includes new two-hour training workshops for these amazing products.

Tools, Tools, Tools — There will be more new and creative software tools to enhance yourMultiValue tool chest plus many new technologies for MultiValue. This year’s program haslots of .NET for sure, but also other important technologies including XML, SOA, RFID,SOAP, SQL, Linux, and Web Services.

Also new this year are special system administration tips and techniques sessions given byMultiValue database providers who are loaning us their systems experts to give you aninside scoop on effective system administration.

And of course, top executives will present their annual corporate updates. Meet key executivesfrom every core provider of the MultiValue database and hear first-hand reports of future prod-uct plans. Participating companies include IBM (UniVerse and UniData), jBASE International(jBASE), Northgate Information Solutions (Reality), ONgroup (ONware), Raining Data (D3,mvBASE, mvENTERPRISE), Revelation Software (OpenInsight), and VIA Systems (UniVision).

There’s lots more happening — including International Spectrum’s Database Challenge —and you can see the whole schedule of events on www.intl-spectrum.com. We hope you’llregister to attend, and we look forward to seeing you in San Diego.

J A N U A R Y / F E B R U A R Y 2 0 0 5

tm

MO/EMentor Pro

Native PICKMULTIVerse

PCVerse PICK OAPower95

R91 REALITY

RPL

AP/PRO

MV.ENTERPRISE

G U S G I O B B IChairman

M O N I C A G I O B B IPresident

N I C H E L L E J O H N S O NEditorial Manager

T E C H N I C A L A D V I S O R Y B O A R DHenry Eggers, Independent Consultant

Tim Holland, Holland Consulting

Bruce Johnson, Tag Consulting

Ronald H. Jones, Ron Jones Consulting

Nathan Rector, Natec Systems

Bryan Shumsky, VIA Systems

Melvin M. Soriano, Eagle Rock Information Systems

M I S S I O N S T A T E M E N T International Spectrummagazine’s editorial mission is to be the premier independent source ofuseful information for users, developers, and resellers of MultiValue data-base management systems, open systems business database solutions,and related hardware, software, and peripherals. Published bimonthly,International Spectrum provides comprehensive coverage of the products,companies, and trends that shape the MultiValue marketplace as well asthe computer industry at large — helping its readers get the most out oftheir business computer systems.

IDBMA’s International Spectrum is published six (6) times per year at thesubscription price of $40.00 U.S. in the U.S.A.; $45.00 U.S. in Cana-da and Mexico; $50.00 U.S. for other countries. Single copy rates are$7.00 U.S. in the U.S.A. and Canada, and $9.00 U.S. in all othercountries. International Spectrum is published by IDBMA, Inc., 311 4thAvenue Suite #513; San Diego, CA 92101; Tel: 619/515-9930; Fax:619/515-9933 E-Mail: [email protected]; Website:http://www.intl-spectrum.com. Copyright 2005 International DatabaseManagement Association, Inc. All rights reserved. Reproduction inwhole or in part, without written permission, is prohibited.

PRINTED IN USA • ART AND DESIGN: CP Design; 858-642-6878;San Diego, CA

NEWS RELEASES/UNSOLICITED ARTICLESInternational Spectrum is eager to print your submissions of up-to-the-minutenews and feature stories complementary to the MultiValue marketplace.Black and white or color photographs are welcome. Although there is noguarantee a submitted article will be published, every article will be con-sidered. International Spectrum retains all reprint rights.

International Spectrum is a registered trademark and MultiValue is a trade-mark of IDBMA Inc. All other registered trademarks and trademarks are theproperty of the respective trademark holders.

MV•BASE

— G U S G I O B B I , C H A I R M A N , I D B M A , I N C . —gus@int l - spec t rum.com


SPECTRUM®24th AnnualSpectrum Show

International Spectrum 2005 MultiValue Conference & Exhibition

February 22-25, 2005Hyatt Islandia Hotel & Marina on San Diego’s Mission Bay

[ F R O M T H E I N S I D E ]

24th AnnualSpectrum Show

Where in the world isVMark, a MultiValue com-pany that many thoughtfell by the wayside yearsago? Attendees at therecent Spectrum show inLondon discovered thatthat certainly wasn’t thecase when they encoun-tered VMark’s exhibitionbooth. It turns out thatVMark is alive and pros-pering, distributing U2products in various partsof the world with head-quarters in Spain, andoffices in France; London;Stuttgart, Germany;Stockholm; and Sao Paulo,Brazil. Many people, particularly in the UnitedStates, had the impressionthat VMark met itsdemise when the compa-ny merged with Unidataand became Ardent in1998. However, theVMark name lived onwhen Pablo Fuertes, whowas then country managerfor VMark Spain, decidedit would be in the bestinterest of his customersto become an independ-ent entity and foundedthe VMark Group.

Headquartered in Madrid, the VMark Groupwas expanded into France in 2002, becom-ing the sole distributor for U2 products.More recently, Fuertes opened an office inManchester in the U.K. “Having opened aU.K. office earlier this year, it was especiallyimportant for VMark to attend this year’sLondon Spectrum show,” said Vicki Barrow,account manager. “It had been apparent fromtalking to U2 customers that it was confusingfor them to hear the name ‘VMark’ again,having thought that the name ceased to existafter the merger with Unidata. This responsewas also true at the Spectrum Vendor Fair —upon seeing the VMark booth, attendeeslooked surprised and puzzled, asking ‘Is thisreally VMark?’ and ‘I didn’t think they existedanymore.’ Attending this event gave us thechance to explain the history of the company

and tell people what we are doing in the U.K.and throughout Europe.”

The VMark Group currently employs over120 people. Eric Delobelle, salesmanager, said VMark’s success asa U2 distributor can be con-tributed to many factors. “Many ofour employees have years of expe-rience within the U2 market, mostbeing ex-employees of Unidata,VMark or Ardent,” he said. “Thishas allowed us to provide highquality services to our customers.Taking the time to employ theright person for a position hasmeant that our staff turnover hasbeen very low, giving customers acontinued service. We have alsoreceived a lot of support fromIBM, and have been able to usethis to benefit our customers.”

In addition to the distribution of U2 prod-ucts, VMark has a policy of working with itscustomers to solve business issues rangingfrom IBM pricing and procedures to defend-ing their investment in U2. Recently, a VMarkbusiness partner was interested in using Red-Back for the new version of their application.Before they made this investment, they wantedassurance from their customers that this wasthe right thing to do, according to Barrow.VMark organized a meeting at IBM regional

headquarters where IBMhelped validate that Red-Back was the best choiceand also confirmed thatthere was a commitment tothe future of the product.

When it comes to partner-ing, VMark puts a lot ofaction behind its words.“VMark works with bothIBM and our customers to

help communicate the posi-tion of U2 as an activelydeveloped and supported

IBM product set,” Delobelle said. “We areable to relay our customers’ needs, their prob-lems and issues to IBM, letting them know

what the U2 market wants, whatthey need, making sure that theirindividual voices are heard withinIBM. We are also in the position tobring IBM to the customer, recentlyworking with IBM to stop thedeployment of SAP at a large end-user site and, along the way, doubletheir U2 license count!”

VMark is also assisting U2 cus-tomers to expand their businessinto Europe, from providing themwith free office space to helpingwith the legal issues. A Frenchbusiness partner wanted to openan office in Spain but wasn’t surehow to do it, Barrow said. VMarkfacilitated by hosting meetings in

Madrid, offering an office in VMark Spain,initiating customer contact, and helping thepartner understand Spanish culture.

In addition to expanding further acrossEurope, VMark is looking to expand into theU.S. in the future, creating a large network ofdistribution offices. VMark believes the keyto success is the ability to offer a flexibleservice that fits around an individual cus-tomer’s needs, providing much more thanjust the supply of products. is


“VMark recently workedwith IBM to stop

the deployment ofSAP at a large

end-user site and,along the way,double their U2license count.”

– ERIC DELOBELLESales Manager

The VMark Group, headquartered in Madrid, Spain, recentlyopened a new office in Manchester, U.K., which joins locationsin France, London, Stuttgart, Stockholm and Sao Paulo.

VMark?Where in the World Is

Early-bird registration discounts now available on www.intl-spectrum.com

ANNOUNCING!International Spectrum’s First Annual

DATABASE CHALLENGE

To find out more information

about participating in the

database challenge email

monica @intl-spectrum.com

to request contest rules

and entry forms.

◆ New half-day training workshops conducted by software authors

◆ Never before presented MultiValue software utilities

◆ The latest in Web-based MultiValue Rapid Application Development tools

◆ .NET and other new MultiValue compatible technologies

◆ Annual corporate product updates by core MultiValue database providers

INTERNATIONAL SPECTRUM 2005 MULTIVALUE CONFERENCE & EXHIBITION

February 22-25, 2005 I Hyatt Islandia Hotel & Marina on San Diego’s Mission Bay I San Diego, California

International Spectrum—

Bringing theInternationalMultiValueCommunityTogether

International Spectrum—

Bringing theInternationalMultiValueCommunityTogetherMark your

calendar for February22-25, 2005 today

and plan to attend theMultiValue event

of the year!

Taking the Bite Out of

IsMultiValue



aMultiValue



In the last couple of years, many MultiValue shops here in the United States

have been thrown into a panic by new federal government mandated IT

controls and audit procedures stemming from legislation known as Sar-

banes-Oxley. I have been amazed and dismayed, amused and bemused watching

the level of wailing and gnashing of teeth in some of these shops when told they

must take the time to do things like track their program change requests, control

who can change live data on the system and keep records of such data manipu-

lation, and restrict programmers from jumping on the production systems and

modifying programs on the fly with no authorization, oversight, or testing. The

strong, sometimes vile, reactions are such that you would think that someone

had just killed their kitten.


W . C L I F T O N O L I V E R , C C P

COBITIT Controls: What Are They and Who Wants Them?

PART 1

10 ◆ INTERNATIONAL SPECTRUM JANUARY/FEBRUARY 200510 ◆ INTERNATIONAL SPECTRUM JANUARY/FEBRUARY 2005

If you suggest that these kinds of con-trols are generally accepted industrybest practices that the Big Kids use andare a Good Thing, you had best bewearing asbestos underwear. Yet, inmany cases the lack of IT controls ispart of what causes corporate IT peo-ple to view MultiValue systems and thepeople who work with them as sec-ond-class citizens, uneducated rubeswho don't know better, “cowboys,” ifyou will. On one IT controls relatedproject, the auditor came right out andtold me, “This [MultiValue] system isour dinosaur. Every time we try to talkto them about controls, they tell usthat [MultiValue] systems can't be con-trolled.” The point was then made thatif this was true, they needed to get ridof it and put in a mainstream systemthat did lend itself to industry bestpractices.

So while complaining that this vendoror that vendor is not doing enough topromote MultiValue as a “real”databasealternative to Oracle, DB2, or SQL Serv-er, these same MultiValue professionalsare driving themselves out of a job bymaking claims that not only should wenot try to play like the big leagues do,but our tool does not allow us to playin that game.

This, of course, is pure poppycock.

Best practices IT controls, also knownas IT governance, is not about theunderlying technology; it is aboutprocesses. It is a way of looking at theInformation Management and Pro-cessing functions as they relate to thebusiness goals and strategies, support-ing the business processes, minimiz-ing risk, maximizing returns on ITinvestments, and identifying newopportunities to use IT to a competi-tive advantage. Upper level businessmanagement has come to realize thatinformation is a core business assetthat needs to be managed just likefacilities or inventory.

“[IT controls are] safeties that allow forbetter preservation of value throughthe management of risks. IT controlsare like the brakes on a car.Not only dothey serve to stop the car and keep itunder control, they enable the driver toactually go faster and still remain safe.”

So let us make the assumption thatyour shop has decided that best prac-tice IT governance is the right thing todo. How do you approach such a largeset of topics without getting boggeddown and making the same mistakeshundreds or thousands of shops havemade before you? One way is to use awell documented,comprehensive,gen-erally accepted best practices frame-work that addresses the entire set ofbusiness IT functions. COBIT (ControlObjectives for Information and RelatedTechnology) is one of the best knownof these frameworks.

Originally released in 1996 by the Infor-mation Systems Audit and Control Foun-dation, COBIT is now in its 3rd editionand is published by the IT GovernanceInstitute. The information is accessiblethrough the Information Systems Auditand Control Association (ISACA). TheirWeb site is at www.asaca.org. Most ofthe COBIT documents are available fordownload at no charge. In addition tothe PDF formatted documents, mem-bers of ISACA can also sign up forCOBIT Online, a very nice Web-basedreference that lets you quickly findinformation about each control objec-tive. (One of my favorite features is theinclusion of “why-do-it” links to explainthe reasons for including various com-ponents.) The subset of COBIT target-ed at small to medium-sized enterprisesis available for purchase from theISACA Bookstore in the publication“COBIT Quickstart.”

The executive summary makes theclaim, “COBIT is relatively small insize.” The operative word here is “rela-tively.” I suppose over 530 pages ofdocumentation would strike you as“relatively small” if you are used to

Continues on page 12

How do you

approach best

practice IT

governance without

getting bogged down

and making the same

mistakes hundreds or

thousands of

shops have made

before you?

So if your IT department does not gov-ern itself now, it will be governed byothers in the future, either throughmanagement edict or governmentalregulation. IT governance is like a cart,and the IT department is like theproverbial dog tied behind it.The dogcan either follow along willing, or itcan lie down and be dragged acrossthe cobblestones. But go along it will.

Although U.S. publicly traded compa-nies are now required by law to havecertain IT controls in place, IT gover-nance is not about meeting regula-tions; it is not an issue limited to oneor two countries; it is just good busi-ness. As George Spafford of the ITProcess Institute points out,

Taking the Bite Out of COBIT Continued from page 8

ROBUST OBJECTS,

Visit us at Spectrum, or at www.InterSystems.com

YOU SHOULD SEETHIS DATABASE

© 2004 InterSystems Corporation. All rights reserved. InterSystems Caché is a registered trademark of InterSystems Corporation. 12-04 RobMultivalue Spect

has become the world’s fastest growing databaseproduct. Caché easily handles the most demand-ing applications with scalability from one to50,000+ users, with modest demands for hard-ware and maintenance.

We are InterSystems – a specialist in datamanagement for over 25 years, providing 24x7 support to 4 million users in 88 countries.

You can see the new Caché, now with Multi-Value technology, at booth #219 at Spectrum.

Only post-relational Caché offers nativeobject, robust SQL and direct access to a powerful multidimensional database – and nowincludes Multivalue technology, with built-insupport for MultiValue Basic, data structuresand queries. This enables MultiValue developersto preserve their existing application invest-ments, plus make use of the latest Web, objectand SQL technologies.

By enabling new levels of high performanceand rapid Web application development, Caché

ROBUST SQL, ANDNOW MULTIVALUE


dealing with government contracts orlegislative volumes. It is interesting tonote that even the ITGI recognizeshow daunting this can be. Quickstart'sdescription admits that “COBIT, in itscomplete form, can be a bit over-whelming.” In my opinion, it is not thevolume that makes it overwhelming, itis the tortured,pseudo-academic, third-party, passive-voice writing style thatpresents itself as an impediment to acomplete comprehension of whichthe reader might be desirous. (Ahem.)

In this series of articles, I would like tointroduce you to the concepts of theCOBIT Framework and its use.After anoverview, we will take one high-levelcontrol objective and drill down to adetailed control objective, examiningits control practices.Then we will con-centrate on the most critical controlobjectives and the COBIT Quickstartsub-set, with comments directly target-ing problem areas in the typical Multi-Value shop. Some of these problemareas will, quite frankly, be more of aproblem of IT department culture thananything technical.

COBIT consists of 34 high-level con-trol objectives that are then brokendown into 318 detailed control objec-tives. Each one of these detailed con-trol objectives has a number ofdocumented control practices. Butdon't panic! (I suppose I could havebeen a bit kinder and warned you tohave smelling salts or some other bot-tle of restorative at hand.) Part of theexercise is to be aware of all of these,but to make decisions as to whichones are suited to your business andhow best to implement them. Thesedecisions are based on how IT needsto support the business functions, noton how easy or difficult it is going tobe. For example, control objectiveAI5.12.2 (see what I meant) speaks tosegregation of function in the installa-tion of programs. If you are a mediumto large company, it is generally accept-ed by IT professionals that developers

should hand programs off to a differ-ent QA person(s) for testing. Imple-mentation into production is done byyet another group, typically Opera-tions. But if you are a small shop withtwo programmers who are the opera-tions people and the help desk, youaren't going to do that. But you maydecide to formalize the process thetwo follow and document it appropri-ately rather than just letting them dowhatever feels right at the time. Itdepends on your situation. This is amanagement decision. (Don't ask thedevelopers.They will always insist theydon't have the time or the luxury tofollow controls. Of course, they alsoinsist they can't do their job withoutthe root or administrator passwords,too. History has proven them wrong.)

IT governance is also an incrementalprocess.No IT department can just boltthe doors and refuse to service newrequests for a year or two while itchanges all of its procedures in onemassive effort.COBIT objective descrip-tions include some indicators to helpyou determine effort versus reward.

A framework can be defined as “a logi-cal structure for classifying and organ-izing complex information.” (Inter-operability Clearinghouse Glossary,www.ichnet.org/glossary.htm). It tellsyou what to consider doing, but nothow to do it. Think of it as a list ofprocesses, each one of which may ormay not be applicable to your organi-zation at this time. Once you deter-mine a particular control objective isappropriate for your company and thepayback is worth the effort and cost,you decide how to achieve it.After all,your company is probably unique.What worked for otherbusinesses may not workfor you. That is whyCOBIT is a framework,not a cookbook.

The COBIT Frameworkstarts at the top by organ-izing IT processes intofour groups, which it

calls “domains.” These domains are:Planning and Organization,Acquisitionand Implementation, Delivery and Sup-port, Monitoring.

Planning and Organization addresseshow the IT function relates to and sup-ports the company's business objec-tives, planning the IT strategy, anddetermining the best IT organizationalstructure, staffing, and infrastructure.

Acquisition and Implementation dealswith finding, developing, or buyingsolutions that support the IT strategyand making sure they are installed,configured, and modified correctly, inline with the defined objectives, andwith the appropriate levels of securityin place.This domain also speaks to thesystem life cycle of change requests,modifications, and testing to reducethe chances that you will break it onceit is in operation.

Delivery and Support concentrates ongetting the services to the users, ensur-ing continuous service, backup andrecovery, disaster planning, security,help desk,and training. In other words,Operations.

Monitoring addresses, well, monitor-ing. Readers familiar with methodolo-gies such as TQM or Six Sigma knowthat it is difficult to improve a processif you do not have defined, measurablecriteria against which to determinehow well you are doing.

Each of these domains contains a num-

ber of high-level control objectives (34

total for all the domains). Just as an

example, the Acquisition and Imple-

mentation Domain has the following

high-level control objectives defined:

AI1 Identify Automated Solutions

AI2 Acquire and Maintain Application Software

AI3 Acquire and Maintain Technology Infrastructure

AI4 Develop and Maintain Procedures

AI5 Install and Accredit Systems

AI6 Manage Change

Taking the Bite Out of COBIT Continued from page 10


For each high-level control objective,you have a Maturity Model, CriticalSuccess Factors, Key Goal Indicators,Key Performance Indicators, AuditingGuideline, and a list of Detailed Con-trol Objectives. The Maturity Modellets you evaluate how you are doingcompared to international standardsand generally accepted industry bestpractices. Critical Success Factors aidin implementing the necessary con-trols. Key Goal Indicators help youdetermine what the target is. Key Per-formance Indicators assist you in meas-uring how close you are to the target.The Auditing Guidelines are primarilyinstructions and recommendations forthe Information System Auditors, bothinternal and external. (If you are stillupset about this “controls thing,” youcan think of this as a pre-release of theenemy's battle plan.)

The Detailed Control Objectives (318total) are the workhorses of COBIT.This is where we get to specifics. For

example, where high-level controlobjective AI6 says “manage changes,”itsdetailed control objectives talk aboutchange request control, risk andimpact assessment, emergency changeprocedures, and so on. Each of thedetailed control objectives has a list ofcontrol practices to provide further

guidance in achieving it. (So the frame-work is a list of lists of lists of lists? It'senough to make a MultiValue program-mer's heart quiver with joy, isn't it?)

In our next visit, we'll take one high-

level control objective and drill down

through its components. is

( 3 0 3 ) 7 5 5 - 1 1 0 2

w w w . f i t z l o n g . c o m

Fitzgerald & Long

•• Resizes both static and dynamic files.

•• Checks file integrity and identifies most errors.

•• Recommends type, modulo andseparation/block size.

•• Performs extensive statistical analysis.

•• Provides after-hours processing.

•• Produces concise reports.Since 1983, we actively support over 120,000 users.

AVAILABLE FOR UNIX, LINUX, AND WINDOWS

ENVIRONMENTS ON ALL MAJORHARDWARE PLATFORMS.

CLIFTON OLIVER brings over three decades of in-depth experience to his

work in the information industry, much of it in the MultiValue market. His

expertise includes project management, technology management, applica-

tion development, data warehousing, and strategic planning for information

services. He was an engineer for Devcom Inc., the original team that

designed and wrote the landmark Prime INFORMATION database system.

He is well-known as an instructor, speaker, and author for both technical and

managerial audiences. These include being an invited lecturer on software

development, project management, and business ethics for such diverse

organizations as the U.S. Department of Commerce, graduate level business

schools, and the Mountbatten Institute in New York.

He also served as the PICK series technical editor for O’Reilly & Associates.

There was a time

when you had to go

looking for news

online. If it changed,

great. But that’s sort

of like heading down

to the newsstand just

to see if anything

new happened. It

was time for home

delivery.

Then a few years back, Netscape intro-duced a concept where news got sentto you if it changed. It was called RichSite Summary, or RSS, and the concepthas moved well past wherever Netscapeitself has gone.

Netscape initially invented the RSS formatfor the various "channels" on its NetscapeNetcenter back in 1999. Non-NetscapeWeb sites began to incorporate the newformat. The format should look familiar tomany of you by now. It’s XML.

Today you can find thousands of RSSfeeds. Intriguingly, RSS might be thebiggest use of XML, even above e-Commerce. Weblog writers, news agencies, governments, government

agencies, and many personal and com-mercial Web sites are now using the for-mat. RSS can be used with softwaretools in Java, PHP, Perl, Python, andother major programming languages.Many RSS viewer software and newsaggregator programs work on the Web,on the desktop, even within e-mailclients. RSS has somewhat become thenew standard for syndication contentand metadata over the Internet.

RSS is pretty much XML used to share

headlines and links to news articles.

With news articles, the entire article isn’t

usually shared, but metadata, or descrip-

tive information, about the article is; the

metadata usually includes a URL, head-

line, and a summary. RSS has become an

important tool for publishers because

these news feeds can be used to syndi-

cate content, and to integrate third-

party content into your site.

Since it is XML, all RSS files must con-

form to the XML 1.0 specification, as

published on the World Wide Web Con-

sortium (W3C) Web site. I’ve spoken

enough about RSS to probably bore

you to tears but let’s just remember that

it’s a flexible way to tag information so

that it looks like HTML.

How can RSS be used? Let’s look at a

typical example of how a company

moves into RSS.

Webonomics 101

Explaining RSS

News That

GrabsYou


• A publisher or company has some content that it wants to share with others.

• It creates an RSS channel for the content.

• In this RSS channel, the companyincludes items for Web pages they wantto share.

• This new RSS channel can be read byremote applications, and converted toheadlines and links.

• These links can be incorporated intoother Web pages, or read in dedicatedreaders.

• People who read those other Websites and readers will see the links, clickon them, and go to the original site ofthe publisher or company.

As I mentioned, headline sharing is the

most common use for RSS. So publish-

ers tend to really like it. But RSS is also

used for other purposes, by non-pub-

lishers. RSS is a very popular format in

the weblog/blogging world. It’s also

used for photo diaries, classified ad list-

ings, corporate job listings, recipes, crit-

ical reviews, and even for tracking the

status of software packages.

In e-Commerce, RSS feeds are used as a

mechanism of information delivery. For

example, Amazon provides customized

RSS feeds based on its Web Services

platform. You can track top or preferred

books in your personal news reader, or

include information on your Web site

about related books for sale at Amazon.

Because of all these uses, RSS hasbecome rather popular in the last fewyears. You can check out sites that listout a bunch of RSS channels, such as

B Y M E L V I N S O R I A N O


Syndic8.com. Yahoo news, the BBC,Amazon, Apple Computer, Wired,Rolling Stone, and CNN are among themany popular sources of RSS feeds.

As the number of RSS channels hasgrown, a new type of software hasemerged: the RSS news reader. Unlikeusenet reads, RSS news readers can bepersonalized more directly. Think ofthem as personal aggregators — theyhelp you find and organize a list ofchannels that you are interested in.Once you have all of your channels, youcan view any of them, using the reader’sstandard interface. The news readerchecks for updates to the channels thatyou are interested in and lets you jumpto URLs in the content.

Looking for more content? You cangoogle for them by typing "filetype:rss"and some topic in a search engine.You’ll get a list of “.rss” files out on theInternet with that topic.

Dedicated search engines make search-ing for content easier. Web site Feed-ster.com monitors zillions of weblogs,and lets you search through an index oflog entries and view them by relevance,date, or blog ranking. Upon submittinga search, Feedster.com automaticallycreates an RSS feed based on yourrequest. This can be added to yournews reader, so that you can see all therecent activity on your search request,without even leaving your news reader.How easy is that?

Daypop.com will search news, blogs,and RSS feeds. It lets you track popularnews within the weblog world. It pro-vides a list of the top 40 most popularweblog URLs. Given that weblogs havecome to the attention of politicians in

the last election, they’ve become quiteimportant to many people. Daypop.comfinds the most commonly linked-to arti-cles throughout the world. It creates alist of top words that are being used inweblogs. Lastly, it also ranks weblogs bycitations, providing an index of weblogsthat are popular with other webloggers.It’s a search engine that essentiallysearches for relevancy. On top of that,you can do custom searches, too. Theranking lists and custom searches, likeFeedster.com, are available as RSSfeeds that you can import into yournews reader.

In the next article, we’ll take a look atthe syntax of the RSS standard and seehow easy it is to keep others updatedwith your information. is

MELVIN M. SORIANO works at Eagle Rock Information Systems (ERIS), an Internet

Application Service Provider and WebWizard/MultiValue Developer. ERIS has deployed

enterprise-wide solutions on most MultiValue platforms and operating systems. HTM-

Mel can be contacted at [email protected] and visited at http://www.eriscorp.com.

You can always call him directly at ERIS's Pasadena, Calif., offices: (626) 535-9658.

In a 20th-year anniversary celebration that began

about one year ago and was inaugurated in the May/June 2004 issue, Spectrum

magazine brings you the last of five installments in our series, “Where Are

They Now?” What appeared to be a daunting research project in the beginning

(“You want me to find how many people who are missing from the industry?”),

turned out to be quite an enjoyable, reminiscent, and often refreshing walk

down memory lane.

In the first installment of this series we took a look at the industry pioneers,

the gentlemen who were brave enough to launch the first Microdata dealer-

ships in the late 1970s (see Spectrum magazine, May/June 2004). The second

and third installments reviewed 25 of the most “influential people” in PICK, as

determined 15 years ago in 1989 (see Spectrum magazine, July/August 2004 for

part 1, and September/October 2004 for part 2). The fourth installment

brought back the who and why of the esteemed “Person of the Year” features,

which appeared on seven covers of Spectrum magazine between 1989 and

1997 (see Spectrum magazine, November/ December 2004). This fifth and last

feature remembers 17 other influential people who made the cover of Spec-

trum magazine between 1984 and 2004.

In the last five issues of Spectrum magazine’s celebration of its 20th year of

publication, we’ve taken a step back into history and remembered a total of 62

people that have made an impact on the PICK/MultiValue marketplace. Of

these, only 11 could not be located. We encourage you to contact us at anytime

in the future ([email protected]), should you know of someone’s

whereabouts that you read about in one of our articles, or you happen to be

just one of those on the Spectrum MIA list.

Where Are They Now?

Where Are They Now?

History of MultiValue

TwoDecades of SpectrumMagazine: InfluentialPeople ThatMade theCover


B y L e a n n e G r e e n


1984 March/April IssueThe inaugural issue of Spectrum magazinefeatured an elderly couple and two youngchildren.The cover story, titled “Computer:Friend or Foe?,” featured this generationalimage as the likely audience that might finda foe in the use of computers.The elderlycouple was actually a pair of modelsbrought in for the photo shoot by the artdirector. Thus, their relationship to the Mul-tiValue market is non-existent, and theiridentities are also unavailable!

The children on the cover,however,do havea personal history with this industry. PaulGiobbi (the 14-year-old young boy) and AlexGiobbi (the two-year-old girl) are the chil-dren of Gus and Monica Giobbi, the publish-ers of International Spectrum magazine.

PAUL GIOBBIon the Cover:Three years before ap-pearing on the cover ofSpectrum, Paul startedhis first job in the PICKmarket, testing dumbterminals at the age of

11 for a Microdata dealership owned byhis father, Gus. He was 14 at the time thecover image was taken in 1984.

“I wasn’t supposed to be on the cover ofSpectrum that day. The fact is that I wasthere to babysit my little sister who wastwo at the time. If I recall correctly, theelderly couple that was modeling for thecover were not providing the desiredimage and for some reason, my parentsgot Alex and me to stand in with them. Iguess that is when the Giobbi nepotismall started.” –Paul Giobbi

Paul Giobbi After the Cover:While still in high school and college, Paulheld various positions for InternationalSpectrum, ending his career there in 1994as VP of Sales. For the next five years heserved as VP of Sales for Jones Business Sys-tems in Houston,Texas. In 2000,Giobbi anda partner bought out the PICK divisionfrom JBS, relocated to Southern California,

mvToolbox I 888-999-7571Contact: John Read, Director of Sales and Marketing

[email protected]

See other examples of how the

MVToolbox programming environment

makes coding fun again at our website:

www.mvtoolbox.comor come see us at the February

Spectrum show in San Diego.

mvToolbox®

putsthe fun back into

MultiValueprogramming again!

Current multi-value utilities make you work

harder than you should. mvToolbox®

breaks programming boundaries that cost

you time and needless effort.

■ Search through multiple files and across

hundreds of programs in seconds rather

than hours.

■ Simultaneously analyze two or more

programs for code differences.

■ Recover effortlessly from programming

errors or system crashes.


and formed Zumasys, a systems integratorfocusing on solutions such as email/con-verged communications (MS Exchange,VOIP from Cisco & Intel); access infra-structure (Terminal Server & Citrix); cus-tom-built servers (HP, IBM and Intel),networking, and security.

Explaining the thrust of his company,Giobbi said:“I’m not out of the MultiValuemarket! It’s like Scarface,’Every time I try toget out, they pull me back in.’ Zumasys stillsupports close to 10,000 D3 users, helpingthem to implement emerging technologysolutions like 3G wireless, which allowsfield personnel to access their MultiValueapplications from anywhere using a cellularInternet connection.

ALEX GIOBBI on the Cover:Alex Giobbi was justtwo years old whenshe made her debut inthe PICK marketplacein 1984,and by the ageof 8, she said she was

proficient in using a PICK system. Sheproudly states that she has witnessed 24years of International Spectrum shows(one of which she attended in her mother’swomb) and said,“I couldn’t be more proudof my family for having created the Spec-trum legacy.”

Alex Giobbi After the Cover:Like her brother before her, Alex heldpositions at the family business through-out her teenage years and during her col-lege summers.Today she works as a LandPurchasing Agent with Pulte Homes Cor-poration out of Orange County. She livesin San Clemente,Calif.,which her mother,Monica, affectionately refers to as “laun-dry distance away.” (Alex explains: “Thismeans it’s too far for me to bring homelaundry on the weekends.”)

1987 Nov/Dec IssueSteve Epstein on the Cover: In 1987 Steve Epstein, a software develop-ment consultant for scientific and engi-neering applications, was chosen to depictthe average computer consultant in thiscover image for a story about the career ofPICK consulting. In a year when most con-sultants were earning between $45 and$75 per hour, Epstein told Spectrum maga-zine that it was time to diverge away fromPICK:“This is probably the last year I’ll beworking exclusively with PICK. By itself, Ifeel PICK simply doesn’t relate to the realworld of computing. There’s no mouse, itisn’t multi-tasking, and the operating sys-tem doesn’t run alongside the ’real’ operat-ing systems of the world.”

Steve Epstein After the Cover: True to his word, Epstein must have left theindustry as we were unable to locate him forthis article.One source,however,emailed usto report that Steve was teaching at MidlandGirls’ School in Los Olivos, Calif. about fouror five years ago. He is no longer there (con-firmed by contacting the school),and no for-warding contact information could beobtained through the school.

1989 May/June IssueMARCIE MILLER on the Cover:It was 1989 and Marcie Miller (then MarcieGebauer, owner of PICK training companyDiscovery Computer Systems), was chosenas the cover image for the story,“Women inComputing.” Depicted as the “average”super woman, super wife, super mom andsuper business owner, Marcie was pictured


Where Are They Now?

Where Are They Now?

Continued from page 17

1985 July/August IssueFRANK BROOS on the Cover:Frank Broos,aka a “supersalesman”in a supermancostume on the cover ofSpectrum, worked withAWA Computers from1972 until 1986 in vari-

ous capacities, starting as technical supportmanager and finishing as deputy generalmanager. He appeared on the cover 10 yearsafter he introduced Microdata REALITY tothe Australian marketplace in 1975. (He wasAWA’s sales manager then, and had becomedeputy general manager at the time of theSuper Salesmen issue in 1985).

“The reason I appeared on the Spectrum mag-azine cover was in recognition of the arguablygreatest penetration of the PICK system in theAustralian marketplace.”–Frank Broos

Frank Broos After the Cover:Broos left AWA in 1986 and subsequentlyestablished C-Itoh Pick Systems in Australia.In 1988 he returned to AWA and took theposition of general manager. In 1989, withthe backing of General Automation andSanderson Plc., Broos led a managementbuyout of AWA and changed the name toGeneral Automation. In 1995, his companybecame a subsidiary of Sanderson Group,and he became the managing director ofthe Australian company. Broos retired fromSanderson in 2000, and since then hasserved as the chairman of Associated Com-puter Solutions, another Australian compa-ny focusing on MultiValue solutions (usingthe UniVerse environment).

(The “baby”Elise is now 17, the youngest ofher three daughters. Elise is pictured withher mom in the updated photo, wearingher soccer uniform. Miller tells Spectrumshe would love to hear from old friends,students or clients, and can be reached [email protected] or 703-476-1418.)

1989 Sept/Oct IssueMIKE ROBERTSon the Cover: When Mike Robertswas chosen to partici-pate in the article aboutcomputer profession-als, he was VP, Informa-tion Systems, with San

Diego,Calif.-based Figi Graphics,a gift com-pany he worked at for 22 years.

“I appeared on the Spectrum cover withthree other people as representatives ofthe different fields in the computer indus-try. We looked at what our job responsi-bilities were in relation to the service weprovided and what issues we each faced.Today that cover would have a lot morefaces on it, representing IT careers andindustries that didn’t even exist then. All Ican say is that it was hard enough to getthe four of us to smile at the same time!”– Mike Roberts

Mike Roberts After the Cover:Roberts stayed with Figi for 22 years, withpositions ranging from data processingmanager to senior VP of Information Sys-tems, and enjoyed the evolution at thiscompany from “soldering terminal wiresto building Web sites.” He left Figi fouryears ago and founded Roberts and Asso-ciates, a consulting business specializingin IT and business operations manage-ment. “For over 20 years I commuted 50miles daily in pretty heavy traffic,”Robertsremembers. “Now, for most days, my com-mute is between my home office and thecoffee machine. The biggest traffic issue istripping over the cats.”


holding her daughter, Elise, who was aboutone and a half years old at the time. Afterleaving McDonnell Douglas just one yearbefore, Marcie had established her ownPICK training company and was well onher way to a long-lasting career as a suc-cessful female entrepreneur.

“I can still remember to this day walking tomy mailbox, opening it up, pulling out theMay/June 1989 issue of Spectrum magazine,and seeing myself on the cover.Even thoughI knew I was going to be on the cover, I wasstill amazed to see myself there! It was trulyone of the most exciting experiences thatI’ve had in my nearly 22 years of being inthe computer industry!” – Marcie Miller

Marcie Miller After the Cover:Miller ran Discovery Computer Systemsuntil April of 2003. Due to the poor econo-my,she explained that the training businesswas slow. She now works full time for Asso-ciation of Trial Lawyers of America (ATLA),where, as a programmer/analyst, she alsoteaches classes and provides technical sup-port to about 180 employees. “I am reallyenjoying my job because of the wide vari-ety of tasks that I am able to do every day.It is fun to work in a building full of peopleafter working virtually by myself for 15years!” Miller said she still runs Discoverypart-time, offering MultiValue and UNIXtraining at customer sites. Continues on page 20


TERRI HALE on the Cover:Although most of hercareer has been spent asan independent contrac-tor,Terri Hale was work-ing for Pick Systemswhen she appeared on

the Spectrum cover for the story about com-puter professionals. She had been broughtonboard by Dick Pick to ensure that themore commonly performed PICK applica-tion tasks became part of the new AdvancedPICK operating system functionality (aka:Open Architecture,D3).

“I can tell you that my husband, althoughexceedingly proud of me [for being on thecover of Spectrum in 1989], was just a tadon the green side, because it wasn’t him.Sorry, Jon!” –Terri Hale

Terri Hale After the Cover:In the late 1980s,Dick Pick “loaned”Hale toJES & Associates to help train the PICKdealers on Advanced PICK. Hale left PickSystems shortly thereafter and joinedforces with JES owner Jonathan Sisk—both

personally and professionally. At JES, shedeveloped and taught AP transition coursesto former R83 users and programmers. Sheand her husband also wrote the first ency-clopedia, from A to Z on AP, named EPICK(with co-author Harvey Rodstein). TodayHale works for Data Systems Group inSacramento, Calif., doing D3 training, pro-gramming, and customer support.

DIRK THAYER on the Cover:When Dirk Thayerappeared on the coverof Spectrum and wasinterviewed for histhoughts on careers incomputing,he was a sys-

tems analyst for Interactive Inc.in San Diego,Calif. The article about Thayer began, “DirkThayer is an unusual computer professional:he is loyal. He has been with one companyfor over five years.” That was in 1989,and 15years later, he is with the same company.

“I can remember wondering why the hecka magazine would choose me to be on itscover, especially since I hadn’t been in theindustry for very long at that time. Thecover did tend to start a number of conver-sations. I would be at some PICK gatheringand someone would just come up saying,’Weren’t you on the cover?’ I met a lot ofinteresting people that way.” –Dirk Thayer

Dirk Thayer After the Cover:Although Thayer has worked for the samecompany for over 20 years, the companyitself has evolved from Interactive Inc. toDataWorks to Epicor Software. His posi-tions have evolved also, and include pro-grammer, PD analyst, customer support,system administrator,pre-sales support, andIS manager.Today he holds the position ofsenior software engineer in the iSolutionsValue Chain group with Epicor. He worksmostly on integration technologies allow-ing the company’s back office products(ERP systems utilizing MultiValue technolo-gy) to communicate with its front officeofferings. (Dirk is pictured here with hisfive-year-old son, Bryce.)

CAROL OLIVER-PARTEEon the Cover:When Carol Oliver-Partee was inter-viewed for the Sept./Oct. 1989 issue ofSpectrum, she wasthe executive VP for

Alert Computer Systems, a software housespecializing in field service applications.She was included in this cover story oncomputer professionals because she hadbegun her career as a documentation spe-cialist, and then evolved into other techni-

Where Are They Now?

Where Are They Now?


INTERNATIONAL SPECTRUM JANUARY/FEBRUARY 2005 ◆ 21Continues on page 22

cal careers, from customer implementationsto customer service to executive manage-ment. By the time the magazine was print-ed, she was working at Pick Systems andabout a year later moved on to The BuckleySchool as systems administrator.

“It was such an honor to appear on thecover! I would like to appear on the coveragain on my Harley. Maybe I could be anambassador,going from state to state on myHarley and letting the world know aboutthe MultiValue industry and most particu-larly about Raining Data and the vision ofDick Pick.” –Carol Oliver-Partee

Carol Oliver-Partee After the Cover:In 1999 Oliver-Partee went back to PickSystems (which then became Raining Data)and simultaneously worked with PetersonResearch designing a GUI front end forOaktree School Management Software.Today she is in Palm Desert, Calif., workingfor Colt Security Investigative Services asCIO and vice president.“I am no longer ina MultiValue environment, and am forcedto use QuickBooks for accounting,” shesaid.“I hate every minute of it!”

1992 Sept/Oct IssueNEILL MILLER on the Cover:In 1992 Neill Miller wasthe chairman of SystemBuilder, an Australiancompany he co-found-ed in 1982 with hisbrother Derek Miller,

who served as the company’s chief techni-cal officer. He appeared on the cover withtwo other Australian businessmen, each ofwhom had a leading 4GL product for PICK-based applications.Miller was quite worthyof this limelight in the early 1990s, havinggrown the business of System Builder intoan international company with OEM con-tracts and a dealer network that spread toalmost every major country in the worldincluding the Soviet Union, which wasquite remarkable at that time.

“I was flattered to be on the cover of themagazine and felt that together with Derek

Miller we had made a significant contribu-tion to spreading the word of the PICK mar-ketplace worldwide. I was proud to beassociated with what was one of the successstories of the IT industry at that time. PICKwas an exciting product and the wholeancillary PICK market was an exhilaratingplace to be. It was littered with exceptionalpeople with vision and an indomitableentrepreneurial spirit.” –Neill Miller

Neill Miller After the Cover:In the mid 1990s,Miller successfuly mergedSystem Builder with Unidata. He played amajor role in the eventual merger of Unida-ta and VMark into what became Ardent Inc.,listed on Nasdaq.Ardent went on to mergewith Informix,which subsequently sold theentire MultiValue business including SystemBuilder to IBM for over one billion dollars.Today System Builder is an IBM-owned andsupported product with hundreds of thou-sands of users worldwide.

After nearly 30 years in the computer busi-ness, Miller explains that, “I needed achange.” Today he spends his time in ven-ture capital investments of start-up compa-nies, and is also a promoter of a substantialportfolio of commercial properties. (Neillis pictured here with his three-year-oldgranddaughter, Sarah.)

BOB GUTHRIEon the Cover:Bob Guthrie, whofounded the success-ful PICK softwaredevelopment housePick Products in 1979,has been a “PICKicon” in Australia. Hebegan his computingcareer in the 1960swhen he was an RPGprogrammer and thenlanded a job withMicrodata in the mid1970s. Shortly beforeappearing on the

cover of the magazine, he had worked toacquire the rights of the Wizard line of appli-cation generators.

“After the death of Ken Simms, the authorof Wizard, I bought the business from hisestate. At the time of the Spectrum cover,three of the world’s leading MultiValue 4GLproducts were being developed in Aus-tralia.This fact prompted the lead article.”– Bob Guthrie

Bob Guthrie After the Cover:About five years after appearing on thecover, Guthrie sold Pick Products to Gener-al Automation and stayed on as generalmanager. In 2003, the Australian assets ofGeneral Automation (by then called GAExpress) were acquired by FusionwareCorp., where Bob has remained in his man-agement position.

JAMES MURRAYon the Cover:James Murray notonly designed anddeveloped the Cue-Bic 4GL and operat-

ing system product, but also managed thebusiness as it stretched from its early suc-cess in Australia in 1984 to huge marketingefforts in the U.S. In 1992, at the time Mur-ray appeared on the cover of the magazine,

CueBic Systems (known in the U.S. asApscore International) was making thenews throughout the industry, as he foughta legal battle with The Ultimate Corp over alawsuit pertaining to exclusive marketingrights of the CueBic 4GL.

“At the time, I remember thinking that it wasa real honor to be on the cover of our indus-try’s most important magazine. Also, Neilland Bob turned out to be nice guys, whichwas a nice surprise.” –James Murray

James Murray After the Cover:Murray is still the CEO of Apscore International.However, over the last year he has founded acompany called LivingMemory.com. Afterspending his time in raising capital for theventure and designing the software behindthe site,Murray hopes for LivingMemory to golive early in the new year. (James is picturedhere with his 10-month-old son,William.)

Excerpts from Spectrum Cover Stories20 Years of Amazing Progress in IT!

1984: “The trend is apparent. In the last year, getting the right information out of a com-puter has been made simpler than ever. For example, a new inquiry system from Microda-ta, called Natural Language, uses a form of artificial intelligence to understand questions putto the computer in ordinary, everyday language. Normally, a question like: “Who beat thesales quota in Florida last year?” would confound most computer systems. But the newcomputer languages will maintain several dictionaries that help it interpret such questions.”(Quote from the Spectrum author of the cover story, “Computers: Friend or Foe?”)

Shucks, we missed the opportunity!

1985: “I think I would like to see the PICK marketplace gain the image and the reputationof IBM. I’d like to see McDonnell Douglas become the IBM of the PICK market; to carry thatsame image of strength and support.” (Quote from Susan Davis, account manager forMcDonnell Douglas in 1985, in an interview for the cover story on “Super Salesmen.”)

1987: “As far as I can see, the economic outlook for the consultant in this [PICK] marketis fantastic! There’s much more business than there are consultants. The market is grow-ing so fast there is an acute shortage of qualified consultants. We’re making more moneythan ever before! (as quoted by an “anonymous longtime consultant” in the cover story,“PICK Consultants: The Glue That Binds It All Together.”)

Anybody looking for a COBOL programmer?

1989: “I see salary studies that compare people that are doing COBOL work and peoplethat are doing PICK work, and PICK programmers are typically much lower paid than some-body with four years of COBOL experience.” (Quote from Kevin King, a financial systemsprogrammer, in an interview for the cover story, “Computer Professionals Speak Out.”)

Harvey’s crystal ball really worked!

1992: “I have seen the future of PICK-based systems and it’s full of icons, mice, bars andbuttons.” (Quote from Harvey Rodstein, author of the cover story, “The PICK 4GL: It WillFizzle Without Sizzle.”)

Guess what? Dick Pick was not Bill Gates (and his son isn’t either).

1993: “Being a ‘Pick’ in the PICK business can often raise some curiosity. Many people,when they get on the phone with me or meet me at a trade show, ask, ’So, are you the Pickof Pick Systems? Aren’t you a bit young? Oh, I get it. Another Bill Gates, eh?’ And I justanswer, ’No. And Pick Systems isn’t just another Microsoft either.” (Quote from Mark Pick,who was with Pick Systems at the time, in an interview for the cover story, “Top 20 Sales-people in PICK.”)

22 ◆ INTERNATIONAL SPECTRUM JANUARY/FEBRUARY 2005 Continues on page 44

Where Are They Now?

Where Are They Now?



My last article,“Legacy ApplicationsGet New Life, WithService OrientedArchitecture,” intro-duced a new ITacronym, SOA (ServiceOriented Architecture),as a way to makeportions of yourapplication availableto business partners.This article discussesanother acronympopularized by Gart-ner, SOBA — ServiceOriented BusinessApplications. Whileat first this mayappear to be nothingmore than buildingWeb-based applica-tions using SOA stan-dards, it is actuallymuch more. SOBA isgiving rise to anoth-er development stan-dard, SODA —Service OrientedDevelopment ofApplications, whichrelies on SOA as wellas building reusablecomponents anddeveloping anddeploying in a dis-tributed fashion.

Why bother if everything is working justfine in your current environment? If thereis no need to take advantage of Web Ser-vices already in existence, remove barriersin doing business with your customers andsuppliers, or attract the next crop of devel-opers then you probably do not need to beconcerned with SOBA. However, if thiscan benefit your organization today or inthe next five years, you may want to con-sider moving in this direction. Let’s take alook at these advantages individually.

You may already be familiar with Web Ser-vices or at least heard of them. There arethousands published that you can use. Allyou need to know is howto ask for informationand what to do with theresults. You do not needto know anything aboutthe platform they run on.You do not need to knowanything about theirdevelopment environ-ment. All you need toknow is how to requestinformation and whatyou want to do with itonce you get it. To get anidea of publicly availableWeb Services (those thathave been published onthe Web for all to use)check the Web sitewww.xmethods.net. Youwill see such services ascurrency exchange, faxservices, SEC filings fromEdgar, worldwide whiteand yellow pages, countrylookup of IP address, etc.

Then there are the pri-vate Web Services. Youmust be invited to usethese and they typicallycome from your cus-tomers and suppliers. Asupplier may invite youto query the status of anorder, track shipping sta-tus, submit an order,obtain technical support, etc. A customermay allow you to check their inventory topermit you to automatically create a replen-ishment order, obtain the phone numberand email address of an employee, checkcorporate news, etc. So what would youlike to invite your customers and suppliersto have access to from your computing sys-tem? Perhaps things like order inquiry,order entry or something that sets youapart from the competition.

In addition, you may find organizationsoffering add-on Web Services to enhanceyour product offering to your customers,

giving you a competitive edge.These enhancements fre-quently require no change toyour existing application. Forexample, you can send yourfinancial statements to a serv-ice that will analyze themwith respect to fraud indica-tors. Or a payables ledger canbe analyzed to help determineareas of savings. Or billingscan be reviewed to help viewreceivables/payments overtime, giving you a differentperspective on your business.Or internal productivityreports can be examined usinganalytical tools to expose hid-den operational trends.

There is little question thatcomputer students are beingexposed to this technology.They will emerge from theirprograms knowing how totake advantage and imple-ment Web Services, SOA,SOBA, XML, etc. New pro-gramming hires in yourorganization will be able towork with your existing devel-opment staff to properly inte-grate your legacy applicationto the world of Web Servicesand SOA, allowing you tooffer SOBA to your suppliersand customers. is

L E E H . B U R S T E I N , founder and presi-

dent of Dynamic Systems Inc., has been a

well known speaker, trainer and consultant

for two decades. He is also co-founder and

executive vice president of Guided Intelli-

gence, LLC, offering SOA solutions for

Forensic Financial Analysis and Opera-

tional Analysis as an ASP. For more infor-

mation, visit www.dynamicsys.com.

If there is noneed to takeadvantage ofWeb Services

already in existence,

remove barriersin doing business

with your customers and

suppliers, orattract the next

crop of developers,then you

probably do notneed to be

concerned withSOBA.

B Y L E E H . B U R S T E I N , P R E S I D E N TDynamic Systems, Inc., Executive Vice President, Guided Intelligence, LLC.

Service Oriented Business Applications:

Dissimilar Applications Working Together

In our last article we introduced the concepts and theory used inthe design of dynamically hashed UniVerse and UniData database

files. Although implemented quite differently, theobjective of these structures, which are generallycalled “dynamic files,” is to expand the size of thefile as data are added in order to prevent the seri-ous performance degradation that can occur whenoverflow is excessive.

This is the sixth article in our series describing database filetheory, structure and maintenance considerations. In all ofour articles we have emphasized the need to minimize over-flow because of the huge impact it has on performance.With the advent of dynamic files in 1987 many Prime

INFORMATION users believed that dynamic files would solve the overflow problem andcompletely eliminate the need to do regular file maintenance. We quoted one of the INFOR-MATION spokesmen as saying, “This is a performance release.” His reference was to Release7.0 of INFORMATION, which was the first release to support the dynamically hashed filestructure. For database administrators who never performed file maintenance it truly was a“performance” release.

You could expect this to be the case with UniVerse and UniData dynamic files as well. Inapplication, dynamic files have eliminated the fearfully long chains of overflow that couldoccur when “statically hashed” files were not resized when needed; but the concept that theyare “maintenance free” is certainly a misconception held by many database administrators.When dynamic files were first introduced, we responded to those who touted the “mainte-nance free” litany with our question, “If dynamic files need no maintenance why are there somany tuning parameters?”

In this article we will discuss the tuning parameters for dynamically hashed UniVerse files. Wewill review the defaults and discuss the theoretical considerations to be considered when tun-ing dynamic files for better performance. Additionally we will describe the information pro-vided by the ANALYZE.FILE utility as well as how and when to use the RESIZE andCONFIGURE.FILE utilities to change file parameters.

Let’s take a look at each of the six UniVerse dynamic files parameters. We will save the dis-cussion of UniData dynamic files for a later article in order to stay within our editor’s sug-gested “file” size.


UniVerse Dynamically

HashedFILES

B Y J E F F F I T Z G E R A L D A N D P E G G Y L O N G

Tuning Parameters


GENERAL — Specifies that the generalhashing algorithm should be used for thedynamic file. GENERAL is the default.

SEQ.NUM — Specifies that a hashingalgorithm suitable for sequential numbersshould be used for the dynamic file. Youshould use this hashing algorithm only forrecords with IDs that are mainly numeric,sequential, and consecutive.

CREATE.FILE and RESIZE allow youto specify the hashing algorithm fordynamic files.

GROUP.SIZE — According to the docu-mentation, static files may have a SEPA-RATION, the size of the buffer in units of512 bytes, between 1 and 8,388,608.Interesting. Creating a file with a separa-tion of 8,388,608 would produce a 4 GIGbuffer. We have never tried to do that! Thedefault separation is 4, a 2,048 byte bufferfor dynamic files. However, dynamic filesdon’t use the term “SEPARATION.”Instead the size of the buffer is called“group size.” The parameterGROUP.SIZE allows the buffer size to beset to 2,048 or 4,096 bytes — a separationof either 4 or 8.

GROUP.SIZE 1 sets the buffer to 2,048bytes and

GROUP.SIZE 2 sets the buffer to 4,096bytes.

Are you wondering why a group size of “1”is 2,048 and a separation of 4 is 2,048? Itmight be just a little something to keepthings interesting because file maintenancecan be a bit on the “dry” side but we believethe real reason is that the design for Uni-Verse dynamic files is based on the PrimeINFORMATION model. There was no

such parameter as SEPARATION inINFORMATION static files. All static filesused a buffer of 2,048 bytes. When dynam-ic files were introduced, the option of usinga larger buffer was added but only fordynamic files. In those days the disk driveswere formatted into “tracks” much like the“tracks” on an LP record (rememberthose?). Each track was subdivided into 9“records” which could hold 2,048 byteseach. The “normal” size of a buffer was onephysical disk record, therefore a group sizeof 2 would mean two physical disk records.

GROUP.SIZE may be specified with theCREATE.FILE and RESIZE utilities. It isNOT available via the CONFIGURE.FILEutility. The on-line HELP for GROUP.SIZEis wrong.

When we create or resize a static file, theparameters to set the hashing algorithm,modulo (number of groups) and separation(number of bytes in each buffer) are speci-fied. When creating a dynamic file, thehashing algorithm is specified by the use ofeither GENERAL or SEQ.NUM. Thebuffer size is specified by GROUP.SIZE{1|2} and the modulo is set via the MINI-MUM.MODULUS keyword. When thesekeywords are not used, the file defaults toGENERAL, GROUP.SIZE 1 and a modu-lo of 1 (one group).

Reminder: Dynamic files are imple-mented as directories containing twofiles, one is the primary data file,called DATA.30, and the other is theoverflow file, called OVER.30.

LARGE.RECORD — The use of theterm “large record” in the UniVerse data-base file context means the data record islarger than the buffer defined by the sepa-

But, before we discuss the parameters weneed to describe how a dynamic file is cre-ated. The UniVerse command CREATE.FILE will prompt for the file name if notprovided on the command line. The nextprompt is for TYPE. If the response iseither the number “30” or the word“DYNAMIC,” a dynamically hashed filewill be created. This is a bit inconsistentwith the use of the meaning of TYPE. Forstatically hashed files the response to thisprompt selects the hashing algorithm,TYPES 2 through 18. However, if thenumber “1” or “19” is entered a directorystructure is created instead of a hashed file.

The RESIZE command can also be usedto convert a statically hashed file to adynamically hashed file. The RESIZEcommand will accept either the number“30” or the word “DYNAMIC.”

Six Dynamic File Tuning Parameters

MINIMUM.MODULUS — If we were creat-ing a statically hashed file with CREATE.FILEwe would be asked to enter the modulo, thenumber of groups. When creating a dynamic filethe modulo will default to 1 unless the MINI-MUM.MODULUS parameter is includedon the command line.

From on-line HELP:

MINIMUM.MODULUS

Specifies the minimum modulo of the file.This keyword takes an integer argumentgreater than one. This value is also the initialvalue of the modulo of the dynamic file.

This parameter can have a significantimpact on performance. But, we will savethat discussion until we have described allsix parameters.

Hashing Algorithms — There are twohashing algorithms that may be used withdynamic files. These algorithms are namedGENERAL and SEQ.NUM for “sequen-tial numeric.” Static files reference thehashing algorithms as TYPE. The hashingalgorithm can be specified when the file iscreated via the use of the words “GENER-AL” or “SEQ.NUM” on the command line.If neither is specified, the file will default tothe GENERAL hashing algorithm.

From on-line HELP:Continues on page 28


ration. For example, if the separation is 4,the buffer is 2,048 bytes in size. If a datarecord is greater than 2048 bytes it is calleda “large record.” It is also stored different-ly than smaller records. Imagine a recordthat is 3,100 bytes long. In a static file,UniVerse would chop off the last 2,048bytes and place that “chunk” in the over-flow area of the file. The first chunk of therecord, the smaller fragment would bewritten into the primary buffer with theoffset to where the buffer size chunk islocated. In effect, this gets the “lumps” outof the primary buffer.

With dynamic files, the user may specifythe number of bytes in what is to be con-sidered a “large record.” But with dynamicfiles, only the record ID and the offset arewritten into the primary file. There is nofragment of data in the primary data file,DATA.30. The entire data record is writ-ten into the OVER.30 file.

The default for LARGE.RECORD is 80%of the buffer size, which is 1,628 bytes fora group size of 2,048 bytes (80% of 2,035bytes, allowing 13 bytes for overhead).LARGE.RECORD may be set usingCREATE.FILE, RESIZE and CONFIG-URE.FILE. It may be set as a number ofbytes or as a percentage. For example:

LARGE.RECORD 1600 would set thelarge record size at 1600 bytes and

LARGE.RECORD 80% would set thelarge record size to 1628 bytes.

SPLIT.LOAD — The ratio of the numberof bytes stored in a file to the capacity ofthe file is called the “load.” For example,imagine a file with one group and the sizeof that group is 2,048 bytes. Let’s use 2,000bytes for this discussion. If we place 1000bytes in the group we could say the groupis 50% full or we could say the load is 50%.

With each addition and deletion to/from afile, UniVerse calculates the load. A param-eter called SPLIT.LOAD defines at whatpoint the file should split. The defaultSPLIT.LOAD is 80%. In our example,80% of 2000 is 1600. When the bytecount in our file is greater than 1600 thegroup will split and a second group will becreated. In theory, half of the data recordswill be moved to the new group. The mod-ulo is now 2. The capacity of the file hasdoubled, let’s call it 4000 bytes (ignore the

additional 96 bytes, they are not allusable). The number of bytes in the fileneeds to exceed 3200 to cause a group tosplit and create another group yielding amodulo 3 file. The load is calculated onthe entire file, not on an individual group.It is common for a dynamic file to havegroups in overflow and the load be lessthan the split factor.

To set or change SPLIT.LOAD, use CRE-ATE.FILE, CONFIGURE.FILE or RESIZE.

MERGE.LOAD — There is anotherparameter called MERGE.LOAD whichcauses the file to decrease in physical sizewhen data are deleted. The default is 50%.Again, the load is calculated on the entirefile, not the individual groups. In ourexample let’s imagine that most of the dataare purged from our modulo 3 file whichhas the capacity of holding 6000 bytes. Ifthere are 2400 bytes of data in the file theload will be 40%. This is less than thedefault MERGE.LOAD factor. In theorythe last group in the file will merge with

ANALYZE.FILE EXAMPLE1 STATS

File name . . . . . . . . . . . . . . EXAMPLE1Pathname . . . . . . . . . . . . . . EXAMPLE1File type. . . . . . . . . . . . . . . DYNAMICHashing Algorithm . . . . . . . . GENERALNo. of groups (modulus) . . . . 18 current ( minimum 1, 0 empty,. . . . . . . . . . . . . . . . . . . . 5 overflowed, 1 badly )

Number of records . . . . . . . . 774Large record size . . . . . . . . . 1628 bytesNumber of large records . . . . 1Group size . . . . . . . . . . . . . 2048 bytesLoad factors . . . . . . . . . . . . 80% (split), 50% (merge) and 79% (actual)Total size . . . . . . . . . . . . . . 55296 bytesTotal size of record data . . . . 29948 bytesTotal size of record IDs . . . . . 5600 bytesUnused space . . . . . . . . . . . 15652 bytesTotal space for records . . . . . 51200 bytes

Number per group ( total of 18 groups )Average Minimum Maximum StdDev

Group buffers. . . . . . . . 1.39 1 4 0.76Record . . . . . . . . . . . 43.00 18 62 12.56Large records. . . . . . . . 0.06 1 1 0.23Data bytes . . . . . . . 1663.78 459 7580 1495.27Record ID bytes . . . . 311.11 129 442 88.69Unused bytes . . . . . . 869.56 168 2000 654.56Total bytes . . . . . . . 2844.44 2048 8192 0.00

Number per record ( total of 774 records )Average Minimum Maximum StdDev

Data bytes . . . . . . . . . 38.69 19 6155 221.27Record ID bytes . . . . . . 7.24 1 25 3.82Total bytes. . . . . . . . . 45.93 20 6180 221.93

HASHED FILES Continued from page 27 the first, meaning data in group 3 willmove to group one and the third groupwill be deleted.

To set or change MERGE.LOAD, useCREATE.FILE, CONFIGURE.FILE orRESIZE.

Changing Parameters

Through changing some of the UniVersedynamic file parameters you can potential-ly reduce overflow, conserve disk space andreduce system overhead, which will resultin faster system performance. That soundsgood. So how do you get started?

We start by looking for opportunitieswhere setting MINIMUM.MODULUSmight be used to reduce unnecessary split-ting and merging. Such processes wherenew dynamic files are created and populat-ed or transaction files are populated andcleared are good cases to consider. Theamount of time saved by pre-allocating thefile can be huge.

ANALYZE FILE EXAMPLE


In a small experiment to quantify the“cost” of splitting we created a dynamic fileusing all of the defaults — a modulo of 1.We wrote a BASIC program which wrote50,000 records of 500 bytes each into thefile. The total elapsed time was 883.46 sec-onds. The file split to a modulo of 15,671.We were the only user on an HP UNIXcomputer.

Next we deleted the file and recreated itusing the MINIMUM.MODULUS key-word and set it to 15,671 which pre-allo-cated the required disk space. We ran thesame program again and it completed in367.15 seconds. The pre-allocated file waspopulated in 40% of the time it had takento populate the file where splitting wasrequired. This is more than 2.4 times fasterand the only difference was the use of theMINIMUM.MODULUS parameter.

For a transaction file that is purged andrepopulated frequently, setting the mini-mum modulus so that the file does notcontract to a modulus of 1 and thenexpand again may improve performance.

The selection of a minimum moduluscould be based upon an average file size forthe life cycle of that particular file.

The second opportunity that can substan-tially reduce overflow in a dynamic file ischanging the hashing algorithm. But, thisis not a simple process given only the toolssupplied by UniVerse. Of course we useour file maintenance utility, FAST c/s, todo this; but here is a technique you can useto determine if testing the other hashingalgorithm is worth the effort.

Which files are updated and/or read con-tinuously by the majority of the users? Ifyou could improve performance for one ortwo processes, which processes would youselect and what files are used by thoseprocesses? Once you have a prioritized listof target files, use the UniVerse command:ANALYZE.FILE filename STATS LPTRto produce a statistical report and send itto the printer. In our example of the filecalled EXAMPLE1, notice the line in thefirst section of the report labeled “No. ofgroups (modulus).” Our file has “0 empty,

5 overflowed, 1 badly.” The 0 (zero) emptyis a good sign. This means that all of the 18groups contain at least one data record.The 5 overflowed means there are 5 out of18 groups that have one overflow buffereach attached. The “1 badly” is a badlyworded and misleading choice of words.When there is a group that has more thanone buffer of overflow attached, ANA-LYZE.FILE will use “badly” to describe it.In this file the largest group is 7580 byteswhich means there are at least 3 buffers ofoverflow attached to one of the groups.That would seem to indicate that at leastfor the one group, performance would bepoor. But, notice the last section of thereport called “Number per records.” This issupposed to list the average, maximum andminimum sizes. Notice the Maximum col-umn reports Data bytes for one record asbeing 6155. This report says that theLARGE.RECORD size is 1628 bytes, thedefault. It also says that there is one “largerecord.” So, we can assume that there is a6155 byte record in the file.



We investigated using a UniVerse query:

We agree that there is ONE recordgreater than 1628 bytes but we do NOTagree that it is 6155 bytes. UniVerseappears to be counting the unused spacein the large records overflow buffers as“data.” Notice that the second largestrecord is only 361 bytes.

SIZE is an I-descriptor that uses the LENfunction on @RECORD. This is veryhandy when looking for large data records.

Our mission is to determine if theSEQ.NUM hashing algorithm will pro-duce better results. If we create another fileusing the SEQ.NUM parameter and copyall of the copy into it, we can then useANALYZE.FILE and compare the statis-tics. We will look at the “overflowed” num-ber to determine which has the smallestnumber of “overflowed” groups as well asthe number of “badly” groups. Next we willlook at the statistics for “Data bytes” pergroup. We want to use the hashing algo-rithm that produces the statistics where theaverage, minimum and maximum are clos-est. In the idea file these three numberswould be equal, meaning that each groupcontains an equal number of bytes.

But we still have not answered the questionas to whether testing the alternate hashingalgorithm is worth the effort. Our tech-nique is to calculate the % standard devia-tion, which simply divides the standarddeviation by the average. But, because wemust distribute records and hope the bytecount per records is homogeneous we willuse the “Number of records” per group datafor our calculation of %SD. To calculate%SD, use the StdDev number (221.27) asthe denominator and the Average (38.69) asthe numerator. 38.69/221.27 = .17 x 100 =17%, the %SD.

Given the statistics for the file with thealternate hashing algorithm we would cal-culate the %SD of data records per group.

The file with the small-est %SD would be thewinner, which wouldmean that there was lessvariance of number ofdata records per group,more even distribution,and the hashing algo-rithm that produced thesmallest variance wouldbe the preferred hashingalgorithm.

For the SPLIT.LOAD, MERGE.LOADand GROUP.SIZE parameters there is nomeaningful statistical analysis we can per-form to give us an indication that weshould make a change. However, there issubstantial theory suggesting when makingchanges is appropriate.

The SPLIT.LOAD parameter defaults to80%. This value works to maintain thephysical size of the file to be less than 80%full. In a number of experiments (by theauthors) it was noted that as the SPLITfactor was increased, throughput timeincreased and as this factor was decreased,throughput time decreased. This behavioris explained when viewed as a function ofoverflow. As the SPLIT factor increased,overflow increased. As the SPLIT factordecreased, overflow decreased.

These results were readily accommodatedin the theoretical model. When the SPLITfactor is low, more space is available ineach group, therefore the probability ofoverflow is decreased. The reverse is alsotrue. When the SPLIT factor is high, lessspace is available in each group for addi-tional data and the probability of overflowwould increase.

It is tempting to lower the SPLIT factoruntil throughput time for a particular taskcould be minimized. There are, however,two other interactions that must be con-sidered. The first is the physical size of thefile. As the SPLIT factor is lowered, thegroups in the file become more sparselypopulated because the number of groupshas increased. The file will occupy morephysical disk space. Second, as the file isspread over more and more disk space,more I/O will be necessary to read theadditional physical space and the amountof data accessed with each “read” will be

smaller. The effect of the increased I/O dueto the increased use of physical space wasobserved during experimentation when theSPLIT factor was set extremely low.

In choosing the SPLIT factor, a compro-mise must be made between oversizing thefile and increasing I/O and the cost ofoverflow. If performance is critical for acertain file and changing the SPLIT factorcan improve performance for a givenprocess, then the change is recommended.However, a larger physical file is slower forprocesses which SELECT, LIST, andSORT the file. For interactive, randomaccess, reducing overflow tends to improveperformance.

The MERGE.LOAD parameter defaultsto 50%. If the SPLIT.LOAD is changed,you should probably consider changingthe MERGE factor also, keeping them sep-arated by 20 to 30 percentage points. Set-ting the SPLIT and MERGE factors highwill tend to increase overflow but will con-serve disk space. A low setting will avoidoverflow but at the expense of large physi-cal files with sparsely populated groups.Setting SPLIT and MERGE too closetogether creates excess overhead wherethere is little opportunity to realize anyperformance gains via the splitting andmerging operations.

The parameter GROUP.SIZE may be setto 1, a 2048 byte buffer or 2, a 4096 bytebuffer. When most of the data records in afile are close to or greater than 1600 bytes,using the larger buffer could possiblyimprove performance. Increasing thebuffer size to match the physical I/O sys-tem is regularly suggested but we have noempirical evidence that this improves per-formance. Once the buffer is available inmemory, UniVerse searches sequentiallywithin the buffer for the specified datarecord. If the buffer is larger, then moresearch time will be needed, assuming thebuffer is close to 80% full.

Remember from our example that theaverage data record size was substantiallyskewed because of the existence of one verylarge record. This is a good example ofwhen NOT to use average record size todecide if the GROUP.SIZE parametershould be changed. If you were to onlylook at that one statistic you mightincrease the group size when there is nogood theoretical reason to do so.GROUP.SIZE cannot be changed withCONFIGURE.FILE.

HASHED FILES Continued from page 29

LIST EXAMPLE1 BY.DSND SIZE TOTAL SIZE;

LIST EXAMPLE1 BY.DSND SIZE TOTAL SIZE 02:05:42pm 08 Dec2004 EXAMPLE1.. Length...

TEST.LIST 4,807SQL.KEYWORD 361

...=========17,387 TOTAL


When dynamic files were first introducedwe were intrigued by the idea that we mightbe able to set LARGE.RECORD smallenough so that ALL of the data recordswould be written into the OVER.30 file.Our reasoning was that only the IDs andthe offsets would be written into theDATA.30 file which could remain verysmall and we would have created a databasewhere the location of all record data wasindexed — the offset would specify the disklocation and we would have a “direct access”file. We thought this would be a great wayto improve performance. We tested this the-ory and it was VERY wrong. Performancewas much slower. Most likely this wascaused by the excessive overhead because itinsured that every “lookup” would have toread two buffers, the primary file buffer andthe overflow file buffer.

CONCLUSIONS — The best perform-ance is achieved through reducing unnec-essary overhead such as disk I/O and CPUusage. Performing desired tasks in themost efficient way produces the bestresults. Dynamic files provide a number ofparameters through which these objectivesmay be pursued.

Reducing overflow improves performancebecause unnecessary disk I/O is eliminat-ed. Through “tuning” the SPLIT.LOADand MERGE.LOAD parameters, experi-menting with both hashing algorithms,and setting LARGE.RECORD optimally,overflow can be minimized. But, careshould be taken when changing theseparameters to not cause the physical size ofthe database to be excessive.

Reducing unnecessary overhead will reduceCPU usage. Setting MINIMUM. MODU-LUS appropriately will eliminate the over-head of splitting and merging that mayoccur with transaction and work files whichare purged and repopulated during someupdate processes. Enlarging GROUP.SIZEcould reduce overhead necessary to look upthe location of large records.

Dynamic files are more complicated andmore accommodating than staticallyhashed files. Through the use of theparameters the database administrator hasmore control over the behavior of the filethan with static files. Dynamic files donot accumulate excessive amounts ofoverflow when neglected. However, thesplitting and merging of dynamic files

insures that there is more overhead asso-ciated with their use. is

Twenty years ago Peggy Long and Jeff

Fitzgerald were running a critical bench-

mark on a top-of-the-line Prime INFORMA-

TION system. The benchmark aborted.

After several hours of detective work they

identified a damaged file. That started a

discussion concerning performance, broken

files and how to check the internal structure

of files after a system crash.

After several weeks of work using Peggy’s

FORTRAN skills and Jeff’s INFO BASIC

knowledge they were confident that they

understood the file internals. This led to a

utility that would quickly scan a file,

report errors and recommend the opti-

mum MODULO and TYPE parameters

needed to RESIZE the file. A year later

they began marketing FAST, which

evolved from this utility.


year; dollar volume totals; trends; dif-ferences; counts; ratios, among manyother variations. These metrics maybe presented in charts, as a singlenumber, as a grid of numbers andfactors, or even as images. MitsWebDashboards support building fullycustomized dashboard objectsthrough the use of ahighly flexible report pro-cessing engine. Thesecustom objects have fullaccess to MITS reportdata and can use cus-tom logic to show highlycondensed sets of infor-mation, such as colorindicators to highlight keymetrics (e.g., if lower thana 10% profit margin,show a red dot). Alldashboard objects canbe links to the underlyingMITS Flash Screensand/or MITS Detail Dis-plays. When linked to aMITS Flash Screen, theuser may then performOLAP analysis with thestandard interactivereporting functionality pro-vided with MITS.

As a proven solution,the MitsWeb Dashboardhas been incorporatedinto The Reynolds andReynolds Company’s (NYSE:REY)new ERA® XT Advanced Reportingsoftware application. This new appli-cation is being included in new andupgraded ERA XT dealer manage-ment systems. More than 10,000 auto-motive retailers use the Reynolds ERAsystem to run their dealerships. Reynoldshas begun its roll out of the ERA XTAdvanced Reporting application, whichincludes MitsWeb Dashboard.

“With the MitsWeb Dashboard,Reynolds ERA XT Advanced Report-ing offers valuable insight into ourdealers’ business performance, givingthem dynamic reporting capabilitiesthat provide a near-real time view ofcritical metrics,” said Ron Lamb, sen-ior marketing director, Platforms andServices for Reynolds and Reynolds.“The dashboard technology savesdealerships time with its easy clickand point interface while deliveringaccess to business critical informationat their fingertips.”

“In its first implementation, MitsWebDashboard was incorporated intothe Reynolds ERA XT AdvancedReporting application, and it hasbeen a resounding success,” saidFred Owen, MITI’s vice presidentand general manager. “MitsWebDashboard is groundbreaking tech-

nology that adds evenmore capabilities to theMITS tool set, trans-forming how informationis presented to the userand making it even eas-ier to take advantage ofcompany data. The factthat Reynolds is utilizingMITS and MitsWebDashboard in its ERAXT software reallyunderscores the bene-fits of this technology.”

MitsWeb Dashboard isavailable as an optionaladdition to MitsWeb.MITS Version 5.3 alsoincludes the followinguser-friendly features:

• Evaluation Hypercubesno longer have to beregenerated whenMitsServer is activated

• String Search added toIdentifier Lookup Feature inSelect Exploration

• Complete enhancement of allHTML forms to support standardCascade Style Sheets for UserInterface Modifications

• Added Firefox Browser Compatibility

• Enhanced MitsWeb SessionSecurity and Support for XP ServicePack 2 (XP2)

• MITS Implementer / Reseller maynow select Features of MitsAdmin tobe available to users

• Enhanced Column Maintenanceand Column Set Maintenance

• Enhanced Column Functions cre-ated on the fly

• Flash Screen Libraries may beflagged as read-only, allowingresellers to distribute and controlstandard flash screen reports withtheir OLAP offering.

For more details about MITS Version5.3 and the new MitsWeb Dash-board, contact a MITS AuthorizedReseller or visit www.mitsonline.com.

GA ServicesReorganizesSales Force to Insure ContinuedGrowth in 2005Company PromotesLong-time Industry Veteran SherwoodKing to Director ofSales and Adds JohnCurl as Director ofMultivendor Services

GA Services, LLC (www.gasllc.com), a premierservices and supportprovider in the opensystems marketplace,has reorganized its sales and marketingorganization to insurecontinued growth overthe next several years.Sherwood King, a 30-year industry veteranwith a background insales, marketing andsales management, was

promoted to Director of Sales. Theexisting sales force will now report toKing, who will continue to focus oncompany growth in traditional opensystem markets. In addition, JohnCurl has joined the company asDirector of Multivendor Services. Hehas over 30 years experience in theservices industry in various sales andexecutive positions. In his new role,Curl will focus on new business devel-opment and general service sales inaddition to multivendor services.

Both sales directors will reportdirectly to George Harris, presidentand CEO of GA Services. “We arevery excited about these changes,and we feel that this will allow us tofocus on new business withoutimpacting our growth in our tradition-al marketplaces,” Harris said. “Thiswill enable GA Services to quicklyaddress the company’s new servicesinitiatives and its quest to become apowerhouse in the expanding multi-vendor services industry.”

To support the anticipated new business growth, Donna Glasgow,Manager of Sales Administration,has increased her staff so that GAServices will continue to provide theexcellent level of customer serviceits existing customers currentlyenjoy. GA Services has also movedinto its new headquarters in Irvine,nearly doubling its warehouse, depot and dispatch areas.

About GA ServicesGA Services, LLC offers seamlessproduct and service solutions tomanage all aspects of informationtechnology. From professional ITconsulting and design, to integrationand implementation, to services

new productsMITI AnnouncesRelease of MITS

Version 5.3, Show casingthe New MitsWeb

Dashboard Reynolds and Reynolds Incorporates MitsWeb Dashboard Technology

Management Information Tools Inc. (MITI),developer of MITS, the leading onlineanalytical processing/business intelli-gence (OLAP/BI) software for the Multi-Value database market, hasannounced the highly anticipated newrelease of its MITS Business IntelligenceSuite. MITS Version 5.3 will be officiallyreleased in January 2005, and willincorporate the new MitsWeb Dash-board option, along with enhancedfunctionality, stability and security.

MITS is a powerful business intelli-gence solution that offers advanceddynamic and interactive data report-ing. Utilizing MITI’s own Hypercubetechnology, MITS takes information ina company’s computer system andmakes it quickly and easily accessibleto decision-makers. Now MITI hastaken the technology a step furtherwith MitsWeb Dashboard.

MitsWeb Dashboard enables users tohave information at their fingertips asthey’ve never had before. It allows theuser to assemble a Dashboard madefrom MitsWeb Dashboard “objects”showing “key metrics” about theirorganization. Residing on a smallamount of space on the computerscreen, MitsWeb Dashboard can beset up to periodically poll for freshinformation that continuously informsthe user rather than requiring them torun a report or take valuable timeresearching issues. Information maybe displayed with an unprecedenteddegree of immediacy.

“The MitsWeb Dashboard enhance-ment brings MITS to a whole newlevel of functionality,” said PatGilbrough, president of MITI. “Theability to make users aware of keymetrics in their specific area ofresponsibility greatly empowers them to be on top of activity in theirorganization. Productivity will increaseas employees are made aware of key activity.”

For example, the MitsWeb Dashboardmay be used to show how manytransactions are being entered todayversus the same date and time last

The MitsWebDashboard maybe used to show

how many transactions arebeing entered

today versus thesame date andtime last year; dollar volumetotals; trends; differences;

counts; ratios,among many

other variations.

new

smak

ers


management, GA Services offerscost-effective solutions that combinemany years of experience withproven abilities in critical areas suchas client, network, server and storage technology. GA Services isheadquartered in Irvine, Calif., withsales and support throughout theUnited States. For more information,visit www.GASLLC.com or call (949)752-6515.

Management Inform ation ToolsDissolves MITS

Reseller A greementW ith IBM

MITI Brings All Activities Relat-ed to Management InformationTool Software (MITS) In-Houseand Expands Services

Management Information Tools Inc.(MITI) announced that it has terminat-ed its agreement with IBM Corp. thatallowed IBM to resell the MITS onlineanalytical processing/business intelli-gence (OLAP/BI) software products.The agreement was in effect fornearly four years. MITI plans to con-solidate all MITS-related activities in-house, enabling it to provide the bestservice to its resellers and customersin the future.

Pat Gilbrough, president of MITI, saidthat although MITI values its closeworking relationship with IBM, thecompany has made a strategic deci-sion in anticipation of the expansionof its services. “We appreciate all thetime and effort that we have spentwith IBM U2 personnel,” Gilbroughstated. “As our MITS product greatlyextends the value companies havein their UniVerse- or UniData-basedsystems, we expect that our relation-ship with IBM U2 will continue in aninformal and jointly beneficial way.”

Since MITI was already handlingmost of the service functions relatedto MITS, customers will see fewchanges as a result of the end of theIBM agreement. MITI will be continu-ing MITS-related activities as they arenow and expanding its services.Those services include:

MITS Worldwide SupportMITI has handled MITS support inNorth America since 2002. AustraliaMITS Support is now established tofocus on support in Australia andNew Zealand.

MITS EngineeringContinual enhancement and development on all MITS softwarecomponents.

MITS Product ActivationMITS Product Activation and Fulfillmentwill continue to be performed by MITI.

MITS TrainingMITI’s training program will continueand be expanded, with regularly

scheduled quarterly classes in theU.S. as well as additional classesnow being held internationally.

MITS MarketingMITI handles all marketing and press-related issues in regard to MITS.

MITS VAR Integration ServicesMITI Professional Services continues togrow and has been actively creatingMITS Hypercubes for resale by VARs.

MITS Application Support for VARsA very successful program in whichMITI helps sell, install and supportMITS systems for VAR customers,MITS Application Support for VARswill continue and be expanded.

MITS VAR InteractionMITI will continue a close workingrelationship with its VARs.

MITS On-Line Product Ordering for VARsThe new MITS Partner On-Line OrderSystem is now in place to expediteMITS VAR Order Processing.

With many businesses looking for acompetitive edge that will enablethem to effectively analyze and usecompany data, business intelligencetools such as MITS, which offersadvanced dynamic and interactivedata reporting, are becoming morein demand. MITS, utilizing MITI’sunique Hypercube technology, is apowerful OLAP solution that takesinformation in a company’s systemand makes it quickly and easilyaccessible to decision-makers. “Thebusiness activity revolving aroundthe MITS product has continued togrow from year to year with 2004being the most successful yet,”Gilbrough said. “We are proud toreport that we have many very suc-cessful MITS resellers throughoutthe world. These resellers havegreatly benefited by offering theMITS product either in addition to, or in conjunction with, their existingproducts. As MITI moves forward,our focus will be to show how other

INTERNATIONAL SPECTRUM NOVEMBER/DECEMBER ◆ 33

IBM U2 / MultiValue VARS can profitby their involvement with the MITSproduct.”

About Management Informa-tion Tools, Inc.Based in Seattle, Wash., ManagementInformation Tools Inc. (www.mitson-line.com) is the developer of theMITS Online Analytical Processing/Business Intelligence system thatincorporates technology allowing thecreation of OLAP Hypercubes and adata warehousing environment on acompany’s existing host computer.MITS systems are firmlyentrenched in a wide range ofindustries: distribution, manufacturing, vehicle dealership,healthcare and services, to name a few. MITS is available on manyMultiValue databases, includingUniData, UniVerse, D3, jBASE, and mvENTERPRISE, and may be ported to other MultiValue databases in the future. ■


Why Linux?Three things motivated Revelation toimplement a Linux version of its software:location, location, location.

Look at our location in the business world— small to medium-sized businesses,where equipment is retained for years,where stability and reliability are prized,where high software cost is intolerable.Look at the types of applications we typi-cally support — back office, point of sale,or factory floor. These are applicationsthat are not tightly bound to MicrosoftOffice; these are the applications with themost pressure and least risk for migrationoff the Microsoft desktop.

Look at our location in the spectrum (punintended) of developers: mainstream prod-

ucts have taken a factoryapproach, dividing tasksinto smaller, more special-ized pieces. This leads tolarge teams of people withsmall sets of skills. Thesefactory teams are cost effec-tive where labor is cheapand flexibility is not impor-tant. These factory teamjobs are now heading to

India and China. Revelation targets craftdevelopers, small teams of smart people.Small teams respond quickly when theyare needed and cost nothing when they aredone; in essence they are the just-in-timesuppliers of the software industry. Revela-tion and the other MultiValue productsoffer a cost and performance mix thatmatches the needs of the craft developer.Linux offers an environment and a philos-ophy that is compatible with the craft pro-grammer. Linux is a natural location fortoday's Revelation developer.

Look at the location of the new developersin application development. They arelooking at Linux, often from outside theU.S. They glaze over when they see any-thing else. As early adopters of the PC,when “real” work was done on big iron,many PC developers received condescend-


[REVELAT ION TECH T IPS ]

Revelation Software recently released

a Linux version of its flagship product,

OpenInsight. In this article we take a

look at that offering. The look, feel, and capa-

bilities of the product are virtually identical

to the Windows version. This article focuses

on why the product will be useful to you.

An Overviewfor Linux

B Y M I K E R U A N E , R E V E L A T I O N S O F T W A R E

OpenInsight


ing comments from establishment work-ers. Look at the PC platform now. TheLinux world is where the early adopters inapplication development gather today.Linux is a natural location to find tomor-row's Revelation developer.

In short, the Linux world is a place of greatopportunity for us.

What's the Big Deal?So we concluded that Linux offers oppor-tunity. This is not new news. MultiValueproducts have been on Unix and nowLinux for some time. The big deal is thatthe desktop environment is growingusable. “Linux runs the Web” has been therallying cry for the Linux community forsome time. The flip side of that coin is thatpeople have been using Web browsers fromtheir Windows desktop because moving toa Linux desktop was too painful. The bigdeal is that the Linux desktops are almostthere. A Microsoft Office power user willnot be satisfied, but the desktop is goodenough for most people. Windows willremain good enough, too. Apple has trans-formed itself to be a cool skin atop Linux.Desktops are becoming commodity soft-ware. There will be no compelling reasonfor users to switch or to stay on any partic-ular platform except for the drag imposedby their legacy software. As a software ven-dor we want to be comfortable in eitherplace. Revelation has long been gearedtoward the PC desktop user. We predict along period of transition as our customersevolve to this commodity desktop environ-ment. Thus, our Linux implementation isdesigned to enable customers to deploy ina pure Windows, pure Linux, or mixedenvironment.

How Does It Work?Java has made the “Write once, run any-where” promise for years. The trick is tocompile your code to platform neutralbytecodes, then write platform specific vir-tual machines to execute the bytecode.Revelation has always been an interpretedenvironment. Thus, moving to Linuxmeant writing a Linux version of our inter-preters. For the non-visual components,

reading records, running programs, exe-cuting queries, etc., this is straightforward.The hard part is the last bit, the GUI run-ning on the desktop. We turned to Main-soft Corporation, which sells VisualMainWin, a set of libraries to map Win-dows API calls onto Unix system calls.These libraries let us use one code base forboth Windows and Linux, and we runmore or less “native” on each. The non-visual pieces seem faster on Linux, whilethe GUI rendering seems slightly sloweron Linux than on Windows. This is partlya reflection of the fact that the WindowsGUI is tightly bound to the kernel, whilethe Linux GUI is managed by an X server.Because we are using the same code basefor both versions, OI for Linux is largelyrunning code that has been tested anddebugged since the first OpenInsightrelease was issued over ten years ago.Because we are using Visual MainWin'slibraries, the code that would have beennew for Revelation is instead runningwithin a mature set of components. Thebottom line is that OpenInsight for Linuxworks well because it sits atop mostlymature code.

So Why Not Go Straightto Java or .Net?If you need to implement in those tech-nologies, go ahead. OpenInsight will sup-port you. We provide JOI, Java forOpenInsight, which builds MultiValuefriendly, ultra-thin-client Java Forms.Alternatively, it is easy enough to imple-ment Web Services via our CGI interface.Web Services consume XML, a delimiteddata format that is almost a cousin to theMultiValue data storage model. If you cando Web Services, then you can do .Net.

However, reaching out to these technolo-gies moves you away from the ultra simple,ultra low-cost implementations, which arethe hallmarks of MultiValue databases.

OpenInsight for Linux offers a MultiValuefriendly way to rapidly deliver a modernlook and feel for your application, withoutpreventing the use of other methods. Youcan cost effectively use OpenInsight to pro-totype in our multi-value friendly GUI

QM, A New Approachto MultiValue

EasyCo.com800-470-2756

Only $119 per user 10 year upgrades just $30

Proven for 4 years Thousands in use

U2/Prime Compatibility D3 extensions

Network Ready Remote files Client/server API

Useful bundled items: AccuTerm + GUI Coyote Web Server

Companion GUI/Web 4GLs OpenTPH $30/user Neucleus $72/user

Supported Platforms: Windows 98, XP Windows NT, 2000, 2003 BSD, Linux

OpenQM GPL on Linux w/ source



If you have seen OpenInsightfor Windows, you know thatthese windows look the samein the Windows version. To aLinux purist, the Windowslook and feel is incorrect.

If you need to satisfy that per-son, it is possible to adjust thelook and feel of the OpenInsight Win-

dows. However, in the world ofthe commodity desktop, theWindows look and feel is com-pletely acceptable. See the Lin-spire version of Linux for a goodexample.

ConclusionFrom the PC to the various Win-dows desktops, and now to theLinux desktop, Revelation Soft-ware has evolved to satisfy thedesktop user. OIL is the latestrefinement of the product,reflecting our belief that the

Linux desktop is maturing. OIL is imple-

before committing to an expensive project ina different technology. Revelation's goal isto make every step a productive step in theright direction, even if it is not the final step.

What Does It Look Like?It looks a lot like OpenInsight for Win-dows. The following screenshots were takenfrom a Red Hat Fedora KDE Desktop:

F I G U R E 1 - The Application Manager

mented in a manner that supports a grad-ual transition of desktops from Windowsto Linux. Equally important is the fact thatRevelation, like all MultiValue products,embeds the small team, low-cost philoso-phy that is at the heart of Linux.

Some of us will be pulled toward Linux bythe opportunities, some of us will bepushed there by customer requirements,and some will commute between Win-dows and Linux for a long time. OpenIn-sight for Linux and/or Windows, builds onthe strength of MultiValued architectureand years of GUI experience to let youmove your application where you want,when you want, without losing that low-cost footprint that you need to survive. is

F I G U R E 2 - A Screen inthe OpenInsight ExamplesApplication

OPENINSIGHT FOR LINUXContinued from page 35


File CreationTo create a file in SB+ you need to go to "Tools > File Create/Delete >". Create a newfile and you will get the file creation screen (Fig 1).

FILE NAME - Enter a unique name for the file. You should make the name mean-ingful and short as it will be referred to frequently. The name should relate to the datawhich will be stored in the file. The maximum recommended length for a file name is20 characters. Do not use a name which contains spaces or the special characters+-/*:<>=\^ or a name which exists in the Master dictionary.

DESCRIPTION - Enter a meaningful description as this will be used in documen-tation which SB+ can generate. It is recommended that the maximum length be 25characters.

NUMBER OF FIELDS - The number of fields for the file should be entered here.This number doesn't have to be exact, but you should try to be reasonably close whenyou enter this. If you are only creating the Data part of the file, enter 0.

NUMBER OF RECORDS - You should enter the approximate number of datarecords which the file will hold. If this will be a dictionary only file, enter 0.

AVERAGE LENGTH OF RECORD - You can calculate this figure by addingtogether the average number of characters for each field defined for the file. This needsto be an approximate figure as the file may be resized later, if necessary.

USE GLO BAL DICTIONA RY -Enter Y if the dictionary definitions for this file areto be maintained via the Global Dictionary, a central repository of definitions. If youenter M, the dictionary for this file will be maintained as a mix of Global and Local Dic-tionary definition items. You should be aware that dictionary definitions created for this


SB+101By Danny Passig

SB+ Basics: Tools, FileCreation and Deletion

This article will deal with

the SB+ tools, file

creation and deletion.

When most files are

created they are divided

into two separate

sections, a data section

and a dictionary section.

The information which

defines the data

characteristics is con-

tained in the dictionary

section and the raw data,

such as phone number

and zip code, is stored in

the data section.

That is it forthis install-ment. Remem-ber Spectrum

2005 is coming up. If you're there, comeand see me at the Natec Systems booth. Iwill also be doing a presentation on LinuxFundamentals. is

Danny Passig is a senior software engineer

at Natec Systems. He has 33 years experi-

ence in the IT field and is currently on the

adjunct faculty at Colorado State University

- Pueblo in the computer science depart-

ment. Danny has done customer software

development and system installations for

various businesses. He has also worked for

IBM as a support engineer for System

Builder. Danny holds a Bachelor's Degree in

Business Administration/Accounting and a

MSCIT from Regis University.

FILE NAME - Enter the name of thefile to be deleted. Use care with thisoption. Using this option will result in theloss of data. It is best to be careful now,instead of having to recover the deleteddata from a backup. Do not delete a file,unless you are sure of the links between thefile and possible screens, reports, menusand other SB+ processes.

DESCRIPTION - The description youentered when you created this file will dis-play here for your information and may beused to help you further identify the fileyou are about to delete.

When you delete a file, you will be giventhe opportunity to delete all the processesthat use this file. As a general rule youshould delete these processes as they willno longer be needed, but make sure thatyou fully understand what you are doingbefore you delete the processes. If youaren't sure, err on the side of caution anddon't delete the processes.

You may enter /FD at any input prompt toenter the file deletion screen.


file within the Global Dictionary may notbe maintained from the Local Dictionaryand vice-versa. Enter N if this file is to be alocal file.

When you develop a new application, thisis usually the first tool which you will use.Files are the heart of your system, so youshould put a lot of thought into theirdesign. A file must be created before youcan use some of the other SB+ tools todevelop it.

File design determines if you will have an ele-gant system — a system which performs well,can be easily maintained and performs welltechnically or an inefficient, inelegant system.

Decisions made during file design will havea direct bearing on the efficiency and flexi-bility of the system you are creating.

At any input prompt you may type /FC toenter the File Creation process.

File DeletionWhen you need to delete a file you created inSB+ go to Tools > File Create/Delete >Delete File for the file deletion screen (Fig. 2).

Bring your Notebook, not your Check Book!

ViságeBoot Camp

Friday 25th Feb, 1-5pm, Garden Room F

PRE BOOKINGS ESSENTIALsales @ stamina.com.au

Take home a FREE copy of Viságeand start transforming your applications beforeyou leave the last session at Spectrum 2005

SB+ 101 Continued from page 37

We have all heard about RFID (Radio Frequency Identification) in one form

or another. Whether you have to implement it because you are a Wal-Mart

or DOD supplier, or you have just seen it in a few articles.

However, most of us don't really know what it can do, how it works, or even

what issues you may have implementing it. Let alone if you should look at

implementing it.

If you do a quick search on Yahoo for "What is RFID," you get about 2.8 mil-

lion hits. This is a lot of information to look through ... and trust me, very

time consuming. Since I'm assuming you don't have the time to do all that

research, here are the basics of RFID.

There are two main parts to an RFID system: the

Transmitter/Reader (Interrogator), and the RFID Tags

(Transponders). In this article, I'm going to focus

on the RFID Tags.

RFID Tags are basically glorified bar codes. Where a bar

code requires line of

sight to read the data,

RFID data is transmitted

to a reader using RF sig-

nals. This allows the tag

to be read without really

having to search for a

label or bar code.

B Y N A T H A N R E C T O R

RFIDWhat Is It?

[T E C H T R E N D S ]

There are two types of tags you can use:active or passive.

Passive tags reflect energy radiated by areader. Meaning they get their power fromthe RF waves striking them. They have nointernal power source of their own. It is alittle thicker than an average sheet ofpaper. Passive tags have an effective rangeof up to 30 feet (10m).

Active tags include an internal powersource and transmitter. Of course, thesecomponents make the tag even larger.Some active tags are the size of a brick.However, they significantly boost the effective operational range of a tag. Active tags can have an effective range of 100+ feet (33+m).

Transmission speed and range is deter-mined by the frequency used, antenna size,power, and interference. Readers can readbetween 10 and 800 tags per second,depending on the equipment and the environment.



Many of us will be using the RFID tags inretail environments, which means we'll beusing tags that supply an ePC. An ePC, orElectronic Product Code, is a standard nam-ing scheme for items or merchandise. It isthe next evolution of the UPC, or UniversalProduct Code, found as a bar code on mostproducts today. The UPC provides a uniqueidentifier for every product. The ePC pro-vides a unique identifier for every item.

Please note, I said that a UPC is for everyproduct, but the ePC is for every item. TheePC allows you to tag every item in a prod-uct line with its own unique code. There areseveral standards for ePC, each allowing youto contain more information, but the mostcommon at this time is the ePC-96 version.

Here is an example of an ePC-96 code:01.115A1D7.28A1E6.421CBA30A

01 - Version of ePC

115A1D7 - Manufacturer Identifier. 28 bits(> 268 million possible manufacturers)

28A1E6 - Product Identifier. 24 bits (> 16million possible products per manufacturer

421CBA30A - Item Serial Number. 36 bits(> 68 billion possible unique items perproduct)

Like UPC codes, the Manufacturer andProduct IDs are distributed and controlledby a central authority called EPCGlobal.You can find more information atwww.EPCGlobal.org.

Now, the question is, "How do I use theRFID tag in my business?" I'll have to leavethat for another article. is

N AT H A N R E C T O R , a regular

contributor to Spectrum, is owner

of Natec Systems, a consulting

firm specializing in D3, AP and

R83 environments and custom

programming. He can be reached

at [email protected] or

http://www.natecsystems.com.

[email protected]

www.TotaLinkcc.com

Visa ■ MasterCard ■ AMEXDiscover ■ Diners ■ JCBDebit Cards ■ Custom

CREDIT CARD INTEGRATION

■ Seamless Integrations into POS,Order Entry and e-Commerce

■ Single or Multiple Merchants■ TCP/IP - Secure Internet■ Single or Multiple Modems■ Your Clearinghouse or Ours■ VAR Partnerships

WEAVER CONSULTINGYour Source for:

• 200+MultiValue UTILITIES!

• HARDWARE & REHOSTING

• SOFTWARE & PROGRAMMING

• ANALYSIS, DESIGN, TRAINING

Authorized Dealer/Partner for:Raining Data

Monolith CorporationAccusoft Enterprises

ZumasysJBSi

And Others

See our catalog, tips, hints & more:www.weaver-consulting.com

or write/call for free catalogEmail: [email protected]

WEAVER CONSULTING4553 Calumet Court

Shingle Springs, CA 95682-9378

530/676-6052 • Fax 530/676-5675

See us at our Spectrum booth or contact us prior to the show to set appointments to visit with us and see our products.

V IS IT OUR WEBSITE AT WWW.OHMWORLD.COM

OHM SYSTEMS INCwww.ohmworld.com I email: [email protected]

1-800-878-0646 USA

Seeking Reseller Relationships in Europe

Product lines of

Current offices in US, Australia and R&D center in India

(reportingwriting tool)

■ Web-centric Application Development Environment WADE

■ Web-centric Application Virtual Environment WAVE

■ OHM Web Manufacturing ERP/MRP application ■ Customer Relationship Management via

24x7 Internet processing CRM


RFID Continued from page 39

Low frequency 125kHz 10 inches (254 mm)

High frequency 13.56 MHz 3 feet (1 meter)

Ultra-high frequency 433 MHz up to 30 feet 868 MHz (<10 meters)902 to 928 MHz2.45 GHz

Electronic Mail Messages


&It is amazing how often I’ve heard the following complaints from users

in different organizations.

■ I never know when my report finished.■ I have no idea whether my report completed successfully.■ When did my report run last?■ Why do I have to ask the administrator if the report finished?■ If only I could send the results to the client easily.

These are common problems when using legacy systems that are notdesigned to generate and send status messages regarding reports. There’sno fast and easy way to know if the reports are finished, if they finished suc-cessfully, or even if the clients were notified of the results. Let’s face it.There are still a tremendous amount of legacy systems out there. These sys-tems have been upgraded to some extent to meet today’s current needs, butthe upgrades have only scratched the surface as far as correcting the prob-lems. These systems were migrated from older versions where the operat-ing system and the RDBMS were one and the same. All the old timers inthis community remember the older versions, including:

■ Microdata■ ADDS (Applied Digital Data Systems)■ Ultimate■ Fujitsu■ Prime Information (Note: One of the first variants of the PICK O/S originally developed in Fortran by Devcon and sold to Prime Computer.)

Just because you have a legacy system doesn’t mean thatyou’re dead in the water.

Legacy systems still exist because:

■ They’re cost-prohibitive to upgrade — the cost associated with replacing and/or redesigning the application is just too expensive.■ They’re time-prohibitive to upgrade — the time needed to review and analyze all the applications may take years of work.

Just because you have a legacy system doesn’t mean that you’re dead in thewater. There are many ways to improve the output from a legacy system. In

B Y R O B E R T C O L T U N

Using BasicPerl to Send


(All questions must be answered. Incomplete forms will not be processed. Complimentary subscriptions are limited to U.S. addresses.)

1. What is your job function/title?

■■ Principal/Owner ■■ Sales/Marketing

■■ President/GM/CEO ■■ Programmer/Analyst

■■ MIS/DP Manager ■■ Purchasing

■■ Controller/Financial ■■ Consultant

■■ VP/Department Head ■■ Other ____________________________

2. Is your company a (check one):

■■ Computer System Supplier ■■ Dealer/OEM/VAR ■■ Software House

■■ Consultant ■■ End User ■■ Other ________________

3. What MultiValue Databases does your company use? (check all that apply)

■■ D3 ■■ Native MultiValue ■■ Reality ■■ Other_____________

■■ jBASE ■■ uniData ■■ UniVerse ■■ uniVision

5. What major business/industry most clearly describes your company?

■■ Accounting ■■ Medical ■■ Direct Marketing ■■ Legal

■■ Banking/Finance ■■ Dental ■■ Construction ■■ Retail

■■ Education ■■ Insurance ■■ Other ____________________________

6. What are your firm’s approximate gross annual sales?

■■ Under $500,000 ■■ $500,000 - $1 million

■■ Over $1 million - $5 million ■■ Over $5 million - $10 million

■■ Over $10 million - $25 million ■■ Over $25 million - $100 million

■■ Over $100 million - $500 million ■■ Over $500 million

IS 1/05

NAME TITLE

COMPANY

ADDRESS

CITY STATE ZIP

PHONE FAX

E-MAIL WEB SITE

SIGNATURE DATE

FREE

SUB

SCRIP

TION Fax this form to

(619) 515-9933and keep

Spectrum magazines coming FREE for 1 year!!


most cases you’re already working on a sys-tem that’s been migrated to either a Unix orWindows environment utilizing new hard-ware. This was the first hurdle to cross, andit’s already been completed for you.

The next hurdle is to be able to improve theflow of information generated with thesemigrated systems. With a little ingenuity,you can accomplish a better flow by utiliz-ing either Basic or Perl, or a combination ofboth, to send electronic messages.

Perl in itself is just another programming lan-guage, like Basic. There are versions of Perlthat run under Unix, Linux and Windows. ThePerl language is free forming, so it allows youto document the program very nicely if youdecide to. I like it because, like Basic, it con-tains a wide variety of instructions.

The more educated you are, the bet-ter off you’ll be.

I personally write most of my scripts todayusing Perl, rather than using Bourne,Korn, C Shell, Sed and Awk. Let’s not for-get that it’s wise to know the other script-ing languages as well. If you work on awide variety of systems, sooner or lateryou’re going to come across scripts writ-ten in different languages from what youuse. With as fast as this industry is grow-ing and changing, the more educated youare the better off you’ll be.

Once I became familiar with Perl runningunder Unix, I was amazed at its flexibility.Using Perl and the Unix job scheduler,cron, I’m able to write scripts that will notonly execute applications but also sendemails confirming if these applicationswere executed, failed, aborted, etc. Thisway I have data available if upper manage-ment or a client complains about missingdata and/or reports. I’d rather be pro-active in addressing these concerns beforethey become real problems.

If you’d like some sample code for sendingelectronic mail messages using Basic and Perl,please visit www.mount-olympus-systems.com.

Note: The Multi-purpose Internet MailExtension (MIME) is the current Internetstandard for sending e-mail messages.This is because it has the capability ofsending information that contains non

USING BASIC AND PERL TO SEND ELECTRONIC MAILMESSAGES Continued from page 41

US-ASCII character sets, enriched text for-mats, images, audio, and video. I havefound that some electronic mail systems areunable to open attachments I’ve createdusing uuencode in conjunction with theUnix Mail command. I found these emailsystems were able to open these sameattachments if they (the attachments) weregenerated with MIME instead. is

R O B E R T R . C O LT U N is one of

the founders and the CEO of Mount

Olympus Systems, Inc., the developer

of Zeus Data Integration™ software.

He has been working in the IT industry

for over 25 years and specializes in

ETL tools. He can be reached at

[email protected].


Look for e-xtra, on the first and the fifteenth of each month.This e-newspaper has timely industry news, hot topic, classifiedads, and things of interest heard on the MultiValue grapevine.

Have you seen it?Everyone’s talking about it, International Spectrum

e-xtra, the new e-newspaper.

®


THE BUSINESS COMPUTER MAGAZINE

SPECTRUMSPECTRUMe-xtra™

• Let us know what you’ve heard on the grapevine that you think we should look into.

• Let us know what’s new at your company that you’d like other people to know!

• Find out how to add e-xtra punch to your International Spectrum Magazine ad with a complementing ad in International

Spectrum’s e-xtra.

• Take advantage of placing an inexpensive classified ad in International Spectrum e-xtra.

Get your FREE subscription at www.int l - spectrum.com

™

Where Are They Now?

Where Are They Now?



myself looking for a new job and ended upat Unidata.” From there he has made manycareer transitions, including positions backat General Automation and back again toPick Systems (renamed Raining Data), afterhis father passed away and the sale of thecompany to private investors.

Earlier this year, Mark Pick formed a newcompany,Highroad Information Technology,along with two partners. Pick, who servesas the company’s president and CEO, saysthe company focus is not primarily on Mul-tiVaue, but specializes in computer systemintegration, maintenance, security, and sup-port solutions that improve the businessprocesses for companies of any size.

tured on a cover of Spectrum magazine, butas Mark Pick explains below, that is probablya good thing:

“When Spectrum magazine came calling inthe early nineties asking me to be on thecover,we all joked around that it was a jinxto be on the cover—similar to the SportsIllustrated jinx (just ask Rich Lauer). Soonafter my appearance on the cover I foundmyself looking for a new job.” –Mark Pick

Mark Pick After the Cover:Mark Pick eventually became VP of Domes-tic Sales for Pick Systems, but said, “Soonafter my promotion and my father’s hiringantics of certain family members, I found

Spectrum Magazine Asks, “What is the Secret to Your Longevity in the PICK/MultiValue Market?”

Paul Giobbi: “Realizing that the value of core business applications written in PICK wasso strong that people would stake their careers on it! And it is a lot of fun working with cus-tomers to protect and preserve these applications and get them to work in new environmentslike the Web, wireless, etc.”

Marcie Miller: “Because I think that the MultiValue database is the best, most flexible, andthe most fun to work with, I have simply made the choice to stay in the market! When I real-ized that I would have to find a full-time job, I called and emailed everyone I knew in the Mul-tiValue marketplace. Unfortunately, not a lot of companies were hiring, so I was lucky to findsuch a good job with nice co-workers and two great bosses!”

Mike Roberts: “It’s no secret that as long as the informational needs and processing require-ments of business keep evolving, there will be a need for my services. My many years of busi-ness and PICK experience have given me the ideal background to provide these services.”

Terri Hale: “Karma. There isn’t anything I’d rather be doing. I’m pretty much a one-trickpony. Fortunately, there has been enough demand for the PICK product and my skill set, tofeed the pony.”

Dirk Thayer: “I think my secret has been flexibility. As indicated, I have held a number ofvaried positions over the years. This has not only allowed me to grow my knowledge base ofboth the MultiValue systems and alternate technologies but has also positioned me well formy current responsibilities utilizing multiple technologies to integrate our various products.”

Bob Guthrie: “I’m still happy to serve the MultiValue market: it has been good for our cus-tomers and I hope we have been good for it.”

James Murray: “My longevity would have to be attributed to having a vision for agroundbreaking and innovative product; commitment to its success; perseverance in theface of an avalanche of dream killers and people that ’can’t’ rather than ’can’ do; the supportof wonderful customers and the fantastic contributions of every team member that everpassed through the Apscore and CueBic Systems doors over the years.”

Catherine Anbil: “The longer I have remained in MultiValue, [the more] I understandthat OHM has been able to rapidly develop changes in features and functionality. We arequick to understand the change in the market and plan ahead of the demand to stay currentand futuristic with our insight to the market of manufacturing, applications and technology.”

Mark Pick: “It’s all in the name . PICK. How could I ever leave it? I tried to leave the PICKmarket briefly during the sale of Pick Systems and was becoming so disgusted with theprogress of the sale. I was pleased when I got a call from Gil Figueroa inquiring about thesale of the company. I found myself back doing what I know best: selling and servicing PICKcustomers. I guess I’m a PICKie for life!”

Bruce Randall: “I enjoyed working at Columbia Ultimate and with all the great people there.I felt I could make a difference in the public sector. I always wanted to be better and learnedmuch about change while there. To me, the MultiValue world has been pretty straightforward(my sales viewpoint) and selling something straightforward has always appealed to me.”

1993 March/April IssueCATHERINE ANBILon the Cover:The 1993 issue of Spectrummagazine featured 20 PICKsalespeople,considered to beleaders of the industry’s salespack. When Catherine Anbilwas selected to this group,she had only been in thePICK industry for three years,

after founding OHM Systems in 1989 withher husband,Suresh Anbil.Obviously,as a co-owner and the company’s vice president atthat time, her sales skills have proven effec-tive as her company enters its 15th year.

“In October of 2003, Gus Giobbi and Iwere discussing the past and I remindedhim that I was one of the individuals on thecover that year.He raised his eyebrows andsaid, ’That was you?’ Obviously we allchange, including you Gus! And after all, Iwas much ‘taller’ then!” –Catherine Anbil

Catherine Anbil After the CoverOHM is still a strong MultiValue applicationprovider for the manufacturing sector ofJIT, Job-Shop, Repetitive, Mix Mode and dis-crete industries. Anbil remains a co-ownerof OHM Systems USA and a shareholder inthe company’s R&D center in Bangalore,India, and most recently, its offshore Aus-tralia offices.Her primary role continues tobe in sales and operations.

MARK PICK on the Cover:The year Spectrum mag-azine highlighted theindustry’s leading sales-people was the yearMark Pick found himselfat Pick Systems as the

national sales manager. (Out of college, hejoined General Automation, a friendly com-petitor to his father’s company,Pick Systems.From there he joined Seattle OS, whichbecame PickBlue, which became Pick Sys-tems.) Surprisingly, Dick Pick was never fea-

A Final Note From the AuthorTo be honest, I agreed to write this year-long series for Gus andMonica Giobbi at Spectrum not only for a fee, but partly out of apersonal curiosity in the history makers of the MultiValue marketand a desire to find some of the people I once dealt with in myown on again/off again relationship with this marketplace. To putmy intimacy with this industry in perspective, and throw a fewqualifications out there for writing this article in the first place,

the Giobbis have asked me to explain my personal history in this eclectic industry.

My father, Larry Walton, was a PICK programmer in the 1970s and remains one to this dayas a semi-retired consultant at the age of 70. It was his friendship and business relation-ship with the Giobbis that landed me my job at Spectrum in 1987, where I worked full-timeand part-time for the next nine years. It was a great job for a young gal; I helped organizethe Spectrum conferences in Australia and England in the late 1980s, spent many days inthe ballrooms and conference rooms of our national and regional conferences around thecountry, and even enjoyed a few years pretending to be a good editor for Spectrum mag-azine until we found a professional editor in Nichelle Johnson!

My personal and brief “where are they now” story is that I left Spectrum in 1996 to pur-sue a teaching credential (which I received) but accidentally fell into a freelance writinggig for General Automation. This blossomed into a full-time marketing consulting busi-ness that has spanned many industries and projects. With the end of this series forSpectrum, I am also officially ending my career in high tech marketing as my husbandand I prepare to open a coffeehouse in our hometown of Fallbrook, Calif.

Thanks to all of you for your help in finding missing people over the last year, for yourinterest in reading the stories of your peers, and for your continued support of theindustry by reading 20 more years of Spectrum magazine and attending the Spectrumconferences. And thus we end this chapter in our Spectrum history books. –LG


Bruce Randall on the Cover:Bruce Randall was selected for this coverstory on leading salespeople in 1993because he had been leading the salesefforts for Columbia Ultimate (CUBS) forfour years. For over 18 years at CUBS, Ran-dall worked in a variety of sales roles,where he made the biggest impact in thegovernment marketplace. Only four yearsafter the article was published, hestepped out of sales to head the new Pub-lic Sector Division of CUBS, and remainedthere as the vice president and generalmanager until he left the company in Feb-ruary of 2004.

“I never thought I would show up on amagazine. Ordinary people like me don’tget on the cover of magazines, so it wasreally a surprise, and flattering to me that itwould actually happen back in 1993.Thatthe possibility exists that it could happenagain, even with a light-hearted slant, stillputs a big grin on my face and my wife andkids think it is great.” –Bruce Randall

Bruce Randall After the Cover:Today Randall is the president of RevQInc., a company he started with a couple tofocus on three areas: First, they are areseller for The Intelitech Group and mar-ket “Acumen! Government” into the publicsector market. Second, they provide consult-

N E E D E D

PICK PROGRAMMERSfor Florida, California, Atlanta New Jersey, Nashville, Denver, N. Carolina, S. Carolina, Texas,

Philadelphia, St. Louis, Michigan, andNew York City

1 - 8 0 0 - 9 4 9 - 5 4 2 3Chuck Okeson, Software Search

770-934-5138 • FAX: 770-939-6410

[email protected]

PICK PROFESSIONALDon’t make the mistake of placing

your career in the hands of just ANYsearch firm. We are professionals whounderstand the importance of CONFI-DENTIALITY and RESPECT of a job

search, and our database of clients isone of the largest in the country.

Unlike the rest, we will work in YOURbest interests’ to help you further your career. Because of our dedication andprofessionalism, we are recognized asthe leaders in the PICK/UniVerse/Unida-

ta placement industry in the Tri-Statearea and throughout the U.S. So if you

are tired of putting yourself at the mercy of the rest.

CALL THE BEST! Contact...Matt Hart

EXECU-SYS, LTD1 Penn Plaza, Suite 3306New York, NY 10119

(212) 967-0505(800) 423-1964 x 302 Fax: (212)947-8593Email: [email protected]

Consultants Needed Nationwide

ing for the public sector in the debt collec-tion arena and do some custom program-ming in PICK.And third, they are getting thefirst release of their debt collection softwarepackage (called “Revenue Results”) sold intothe smaller government market.

Mike Sutkowski on the Cover:Mike Sutkowski was an obvious choice forthe cover image of Spectrum’s “Top 20Salespeople in PICK,”as he was responsiblefor VMark’s largest revenue-producingregion in North America from 1988 untilthe article was written in 1993. The sameyear he appeared on the cover of Spectrumhe was promoted to vice president of Salesfor VMark Software.

Mike Sutkowski After the Cover:We were successful in tracking down MikeSutkowski and found him at Massachu-setts-based Sepaton Inc., where he holdsthe position of vice president, OEM Sales,for its data protection products.We learnedthat after leaving VMark, he was generalmanager for Data General’s AViiON serverchannels and healthcare divisions. Next, hemoved to Auspex Systems, where heworked as vice president of Sales, EasternRegion. Unfortunately, Sutkowski’s sched-ule was such that an interview could not benailed down in time for our deadline.

Stacey Smith on the Cover:In 1993 Stacey Smith was pictured on thecover of Spectrum magazine and featuredas one of only two women in the articleabout the top 20 salespeople in PICK. Atthe time, she was an “automation special-ist” in the sales division for Dynix Inc., theprovider of one of the leading libraryautomation applications in the country,and had worked for Dynix for four years.

Stacey Smith After the Cover:All of our research methods proved to beunsuccessful in locating Stacey. is

code. When programs are written orresearched, the programmer very oftenneeds to access and view several programsconcurrently.

Wouldn’t it make more sense to seamlesslytraverse data and multiple program recordsfrom within one edit session — withoutthe need for multiple searches and lists tolocate code or data? mvToolbox has creat-ed just such a programming environment.This MultiValue database solution will saveprogrammers tremendous time and effort.

To explain how mvToolbox works, imag-ine that you are standing at the edge of avast forest called Forest A. There are cer-tain “problem” trees that need to be locat-ed and you will have to walk throughthousands of trees to locate the trees youneed to fix. Just over the hill is another for-est, Forest B, that contains a number oftrees that have the same problem. Withconventional MultiValue programmingcapabilities you must complete your workin Forest A, then travel to the next locationand start the process all over again.mvToolbox provides you with innovativesearch tools and the capability to searchboth locations at the same time. Bro Cope,president of mvToolbox explains, “Utiliz-ing the powerful search capabilities ofmvToolbox, programmers can do in min-utes what used to take days or weeks to do.”

With mvToolbox, it is possible to concur-rently search across several accounts, each

mvToolChest


Over the years,programmers

have faced barriersthat prevent accessbeyond the currentrecord, resulting inunnecessary, andtime consuming,repetitive actions.Most programmersworking with Mul-tiValue databaseshave found it neces-sary to take out timefrom their higherpriority program-ming duties todevelop tools oftheir own in anattempt to eliminateat least some of theconstant backtrack-ing and repetition.

But few, if any, MultiValue database pro-grammers have the time and resourcesavailable to really address the major barri-ers that hinder MultiValue database pro-ductivity.

For example, standard MultiValue editorsallow editing only one program record ata time. This results in repetitive edit ses-sions, moving in and out of programsseeking answers to question after ques-tion, and mentally attempting to join dis-continuous and disjointed sections of


The basic programming

utilities supplied with

MultiValue databases

have not been significantly

improved for over

three decades. All of

that changes with the

new mvToolbox

programming environment

with comprehensive

MultiValue programming

tools that take the

problems out of problem

solving.

Reduces Repetition and Increases Productivity

mvToolboxMakes Programmers’ Tasks Easy and Efficient


with multiple files that can contain thou-sands of program records with thousands ofcode lines. The results are then viewed inwork windows that provide functionality tofurther drill down to specific code lines.Editing these lines is a simple mouse clickaway. Searches take seconds or minutes —not days or weeks.

Over 450 tools and functions are integrat-ed within mvToolbox to make the pro-grammer’s day-to-day activities easier andmore efficient. Programmers work in afully mouse-enabled GUI-like environ-ment. This enables programmers to utilizemouse clicks to move through code anddata. A built-in text editor makes changingor adding code a breeze. “Our program-mers, when working solely on code gener-ation, save 20% of their time,” said onemanager in San Diego, Calif.

There is no need to exit mvToolbox, as allsearch and edit functions are integratedand available at any time. Information thatis the result of one command can be usedas input to another command. A program-

mer can go many levels deep within pro-grams and files and return to where hestarted with simple mouse clicks. The codeis not just text being edited, it is viewed bythe mvToolbox as a map, a source of infor-mation to provide easy mobility in andthrough the code.

There is one feature that is worth itsweight in gold. mvToolbox has four levelsof error and disaster recovery. These fourlevels progress from recovering from cod-ing mistakes to full disaster recovery due toserver crashes and power failures. Revisioncontrols also allow records and files to belocked and unlocked by management toprevent unauthorized changes. Revisioncontrols also keep a permanent audit trailof changes to programs.

Another helpful set of tools and functionsenables reconciling current and previousversions of programs or data. mvToolboxhas the power to reconcile up to six concur-rent versions and allow changes or transfersbetween each version. “What I expected to

take eight hours to do, I did in less than onehour,” said a user who needed to reconcile dif-ferences in two program records containingover 15,000 lines of code.

Whether you program daily in MultiValuedatabases or are an integrator making aMultiValue database compatible with otherdatabase platforms, mvToolbox pays foritself many times over. And, in the words ofone veteran user of mvToolbox, “MultiVal-ue programming is fun again!” is

For more information about the power

of this new programming environ-

ment, go to www.mvtoolbox.com or

call 888-999-7571 Ext.102. Their e-mail

address is [email protected].

SJ+ Systems Associateswww.sjplus.com I 9 5 4 • 7 9 6 • 9 8 6 8

That’s one way to keep your data squeaky clean.

The other is with PRC software configuration managementtool as the framework. You gain immediate control and clarity. Then, scale into best-practices that keep you — andyour metadata — clean, clear and credible.

With data trails, defined emergency access, change control,process documentation, check out/in and automated deployment and rollback, you’ll get the control and trans-parency that management and the auditors demand.

They’ll be showering you with praise.

Documents

INSIDE! $7.00 U.S. SPECTRUM INTERNA TI ONAL · Is $7.00 U.S. ® INTERNA TI ONAL THE BUSINESS COMPUTER MAGAZINE JANUARY/FEBRUARY 2005 • AN IDBMA, INC. PUBLICATION SPECTRUM INSIDE!