A Methodology of a CASE-Tool Assessment - uni … · A Methodology of a CASE-Tool Assessment ......
Transcript of A Methodology of a CASE-Tool Assessment - uni … · A Methodology of a CASE-Tool Assessment ......
© 1996 University of Cologne, Chair for Business Computing, Albertus-Magnus-Platz, D-50923 Köln, Germany
Andreas Hierholzer, Georg Herzwurm
A Methodology of a CASE-Tool Assessment
W. Mellis, G. Herzwurm, D. Stelzer (Hrsg.)Studien zur Systementwicklung, Band 8, 1996
2 A Methodology of a CASE-Tool Assessment
© 1996 University of Cologne, Chair for Business Computing, Albertus-Magnus-Platz, D-50923 Köln, Germany
Titelaufnahme:Hierholzer, Andreas; Herzwurm, Georg:A Methodology of a CASE-Tool Assessment / Hierholzer, Andreas; Herzwurm, Georg - Köln: 1996
(Reihe: Studien zur Systementwicklung; Bd. 8ISSN 0944-6605)
Die in diesem Buch genannten Produktnamen sind in der Regel geschützte oder eingetragene Warenzeichen, ohne daßdies ausdrücklich gekennzeichnet wurde.
Bei der Erstellung und Korrektur des Beitrags sowie bei der Endredaktion wurde mit größter Sorgfalt gearbeitet.Trotzdem können Fehler nicht vollständig ausgeschlossen werden. Herausgeber und Autoren können für fehlerhafteAngaben und deren Folgen weder eine juristische Verantwortung noch irgendeine Haftung übernehmen.
Das Werk einschließlich aller seiner Teile ist urheberrechtlich geschützt. Jede Verwertung außerhalb der engen Gren-zen des Urheberrechtsgesetzes ist ohne Zustimmung des Herausgebers und der Autoren unzulässig und strafbar. Dasgilt insbesondere für Vervielfältigungen, Übersetzungen, Mikroverfilmungen und die Einspeicherung und Verarbei-tung in elektronischen Systemen.
Printed in Germany
ISSN 0944-6605
A Methodology of a CASE-Tool Assessment 3
© 1996 University of Cologne, Chair for Business Computing, Albertus-Magnus-Platz, D-50923 Köln, Germany
Table of contents
1 Introduction 5
2 CASE-Tool market-analyses in literature 5
3 Methodological approach for a CASE-Tool Assessment 6
3.1 Aim of examination 6
3.2 Sequence of the examination 6
3.3 Selection of examined CASE-Tools 7
3.4 Survey and data acquisition 8
3.4.1 Catalogue of criteria 8
3.4.2 Evaluation plan 8
4 Results of the CASE-Tool Assessment 9
4.1 Summary 9
4.2 Tendency statements 13
4.2.1 Tendency statements by hardware/OS environment 14
4.2.2 Tendency statements by development methodology 15
4.3 Evaluation of the CASE-Tools using benefit-value-analysis 17
5 Conclusion 19
6 References 21
4 A Methodology of a CASE-Tool Assessment
© 1996 University of Cologne, Chair for Business Computing, Albertus-Magnus-Platz, D-50923 Köln, Germany
Wir danken den Mitarbeitern der in dieser Studie berücksichtigten Unternehmen, die uns bereitwil ligihre Produkte zur Verfügung stellten und Auskunft erteilten, sowie den Mitarbeitern des Lehrstuhlsfür Wirtschaftsinformatik, Systementwicklung der Universität zu Köln, insbesondere Frau AndreaSchütz, bei der Abfassung der Studie behilflich waren.
A Methodology of a CASE-Tool Assessment 5
© 1996 University of Cologne, Chair for Business Computing, Albertus-Magnus-Platz, D-50923 Köln, Germany
1 Introduction
CASE may be defined as a partial discipline of computer science, that deals with the automation and
computer-aided support of development, use and maintenance of software products. With the inte-
gration of methods and procedures, as well as their automation, a computer-aided system shall be
supplied, that ideally supports all phases and activities of the entire system life cycle.
After the first CASE-Tools have been offered to the market in the early 80´s, which did not lead to
the desired success, a new generation of CASE-Tools now tries to contribute to the overcome of the
much-cited software crisis. The conventional tools of the 80´s, that have been designed for mainfra-
mes, COBOL-applications and structured methods, are now confronted with new products for cli-
ent-server-applications, object-oriented methods and object-oriented languages.
Due to the dyamic of the CASE-market and the complexity of products, choosing a CASE-product
in practice is really a hard thing to do. The following article therefore tries to offer a guidance for the
practician, that reflects the state-of-the-art of conventional and object-oriented CASE-technology.
2 CASE-Tool market-analyses in literature
Certainly, there already exists a number of publications [e. g.: Balz93, Gane90], that give an over-
view of CASE-Tools, but these are rather descriptive and they only contain very few critical and
comparative analyses. The studies of various management consultancies are on one hand partially
antiquated and have not been actualized any more. 1 On the other hand, these analyses have been
realized rather in an intuitive and subjective way instead of being scientific and precise. To our opini-
on, a chair of business computing as a neutral institution is able to word requirements regarding
CASE-Tools more independently, because it is not committed to certain suppliers or customers.
1 This is not valid for [Ovum93].
6 A Methodology of a CASE-Tool Assessment
© 1996 University of Cologne, Chair for Business Computing, Albertus-Magnus-Platz, D-50923 Köln, Germany
3 Methodological approach for a CASE-Tool Assessment
3.1 Aim of examination
From the practician´s point of view, a ranking of the „best“ CASE-Tools certainly would be a desi-
rable examination result. But the determination of an ordinal sequence of CASE-Tools is not general-
ly applicable.1
The adequancy of a CASE-Tool for a certain company depends on its strategic aims, the implemen-
ted hardware and software for development and production, the companies´ standards, the structure
of organisation and employees, etc., and is therefore specific for each company.
So this article aims to present general statements concerning the performance of the examined
CASE-Tools. A conclusive evaluation of CASE-Tools can only be done including the specific requi-
rements of each company.
Quality standards like ISO 9000 gain increasing importance for development processes, so that the
requirements that can be derived from these standards are especially interesting. An example for such
a conclusive evalutation is shown presenting two case studies. The weighting of the considered eva-
luation criteria was fixed either by a german insurance company or has been derived from the ISO
9000.
3.2 Sequence of the examination
An advance questioning of manufacturers and suppliers has been realized (the entirety of 60) to gain
an overall view of the CASE-Tools offered in Germany and to receive a mean for the selection of
CASE-Tools we wanted to examine in greater detail. From this entirety we selected in the end 17
manufacturers that fulfilled certain criteria (see section 3.3). We sent them an extensive catalogue of
criteria that allowed a technical and economical characterization of their products.
Within the next step the selected 17 CASE-Tools had to submit to the evaluation, that was led by
our chair. The evaluation was done by presenting the following exercise. Using the CASE-Tools we
solved a system development problem that has been derived from the recommendations of a german
insurance company. The advantage of this exercise was its extent as well as its representativity. Due
1 A benefit-value-analysis for a special CASE-Tool selection contains for example [BeHe92].
A Methodology of a CASE-Tool Assessment 7
© 1996 University of Cologne, Chair for Business Computing, Albertus-Magnus-Platz, D-50923 Köln, Germany
to its extent, the CASE-Tools could be evaluated for their concrete effectiveness throughout all the
different phases of development. The exercise was on the other hand representative for a classical
economic data processing problem, because it may occur not only in insurance companies, but as
well in the services sectors (banks, public authorities etc.) or administrative sectors (Financial ac-
counting, personnel) of other lines of business. The essential part of the exercise was to design a data
base application using modern presentation techniques (ERM, SA, OOA, OOD, etc.) and to trans-
form it into executable code (SQL, C, COBOL, etc.), including different user interfaces (GUI, CUI).
The mentioned catalogue of criteria that has been filled out by the manufacturers served as the me-
thodical base for the comparison of performance of the tools. The implementation of the system de-
velopment problem was lateron used to verify the specifications stated by the manufacturers. Invol-
ving the manufacturers, we gained completeness and the fact that the chair made the evaluation con-
tributed to a sufficient objectivity regarding the performance test of the CASE-Tools.
3.3 Selection of examined CASE-Tools
The examined tools had to show the following characteristic features to be distinguished from other
tools for system development. In this way, they could be classified as CASE-Tools in the sense of
our examination.
• support of several phases of the system life cycle
• the analysing phase belongs to these supported phases
• a common user and data interface integrates the single tool components
• graphical description tools are provided
Furthermore, only those CASE-Tools seemed to be suitable for the purpose of this study, that are
distributed and supported in Germany. 1 So we first of all could limit the entirety of tools to 60
CASE-Tools for this examination. For the reason of capacity we further had to select from these 60
1 An actual market survey for the U.S.A. contains for example [Harm94a] und [Harm94b].
8 A Methodology of a CASE-Tool Assessment
© 1996 University of Cologne, Chair for Business Computing, Albertus-Magnus-Platz, D-50923 Köln, Germany
tools another 17 conventional and object-oriented tools for our case study, that presented the largest
numbers of installations.1
3.4 Survey and data acquisition
3.4.1 Catalogue of criteria
The CASE-Tool evaluation is based on a catalogue of criteria. This catalogue is founded on the ex-
periences that have been made 1991 during the selection of CASE-Tools for education in business
computing at the University of Cologne. The choice was made in cooperation with industrial com-
panies. Due to numerous concrete suggestions, the catalogue has been revised and actualized, paying
special attention to „new“ technologies, as for example the object- oriented system development.
The result is a check-list, containing 400 criteria that focuses on the support of methods (supply of
means for description and proceedings, testing mechanisms etc.).
3.4.2 Evaluation plan
The assessment and comparison of performance of different CASE-Tools is a difficult thing to do.
To achieve a most objective comparison level, you need a plan for evaluating CASE-Tools, that
must be applied consistantly throughout the whole examination. If you want to guarantee the ability
to compare the results, you need to fix the evaluation plan in writing. This is even more important, if
the examination is being led by a team of seven persons, as it was in this special case. The evaluation
plan describes, what the examiner has to be aware of, how to interpret the results and how to com-
ment on them. The most important result of this analysis is to provide consistantly structured evalua-
tion reports for each CASE-Tool.
1 First we intended an evaluation of 20 CASE-Tools. Following the criterion „Number of installations“, we wouldhave had to evaluate the products of Bachman (Bachman Product Set), Andersen Consulting (Foundation) andOracle (Oracle*CASE), too. But these products have not been placed to our proposal at the right time.
A Methodology of a CASE-Tool Assessment 9
© 1996 University of Cologne, Chair for Business Computing, Albertus-Magnus-Platz, D-50923 Köln, Germany
4 Results of the CASE-Tool Assessment
4.1 Summary
Figure 1 shows the range of performance of the 17 evaluated CASE-Tools. A complete list of all the
strong and weak points of the examined tools is not possible within the scope of an article like this.1
An overview of the essential strong and weak points of the evaluated CASE-Tools shows figure 2.
1 The complete documentation of the examination results in [Herz94a] encloses about 2.500 pages.
10 A Methodology of a CASE-Tool Assessment
© 1996 University of Cologne, Chair for Business Computing, Albertus-Magnus-Platz, D-50923 Köln, Germany
Product Manufacturer
Operating Systems Graphical means ofdescription
Methods Special ap-proaches
Generation PM
CM
DOS
MS-Win-dows
OS/2
UNIX
Others DB-Schemata
Code
form-based UI
graphicalUI
ADW Knowledge Ware • • • M-Spec, ERD, AD, DFD, SC SA, SD, IE • • • •case4/0 microTool • ERD ERM, SA, SD • • •DDB-CASE Delta • • • DFD, ERD, M-Spec, FT ERM, SA Merise, SDMS • •Excelerator II Intersolv • • DFD, ERD, C-Spec ERM, RT, SA,
SC • •IEF TI Information
Engineering • • • MVS AD, SC, ELH, ERD, DDD ERM IE • • • • •Innovator MID • • • • • DFD, SERM, SC,
M-SpecERM, SA, SD,RT • • •
MAESTRO II Softlab • • • ERD, FT, Fctfld ERM V-Modell, SETec,Merise, SSADM • • • • •
ObjectTool Rösch Consulting • • • CD, STD OAOD •OM-Tool IQProducts • • • CD, STD OMT •NEW (PREDICTCASE)
Software AG • DFD, ERD, BFD, CFD ERM, SA ISOTEC •Legend:
• : available IE: Information Engineering PM: Project ManagementAD: Action Diagram CD: Class Diagram RT: Real-TimeBFD: Business Function Diagram CFD: Control Flow Diagram SA: Structured AnalysisDDD: Dialog Design Diagram CM: Configuration Management SC: Structure ChartsDFD: Data Flow Diagram C-Spec: Control-Specification SD: Struktured DesignELH: Entity Life History M-Spec: Mini-Specification SERM: Structured ERMERD: Entity-Relationship-Diagram MS: Module Structure SOM: Semantic Object ModelERM: Entity-Relationship-Modeling OAOD: OO Analysis/Design (Coad/Yourdon) STD: State Transition DiagramFT: Function Tree OOSA: OO System Analysis (Shlaer/Mellor) POM. Process Object ModellFctfld: Function Flow Diagramm OMT: Object Modelling Technique (Rumbaugh) VSOM: Vorgehensmodell für semantisches Objektmodell
Figure 1: Range of performance of the evaluated CASE-Tools (Part 1)
A Methodology of a CASE-Tool Assessment
© 1996 University of Cologne, Chair for Business Computing, Albertus-Magnus-Platz, D-50923 Köln, Germany
Product Manufacturer
Operating Systems Graphical means ofdescription
Methods Special ap-proaches
Generation PM
CM
DOS
MS-Win-dows
OS/2
UNIX
Others DB-Schemata
Code
form-basedUI
graphicalUI
ProMod+ debis Systemhaus • • • VMS DFD, ERD, FT, C-Spec, M-Spec,RT, SC
SA, SD, ERM
Software through Pictu-res
IDE • AD, DFD, ERM, CD, SC, STD SA, OMT •
SDW SDW Software • • DFD, ERD, SC ERM, SA, SD • •SOM-CASE Uni Bamberg • • • SOM, POM VSOM •Systems Engineer LBMS • DFD, ELH, ERD, MS SA, SD, ERM LBMS-SE • • • • • •Teamwork CADRE Technolo-
gies • ERD, M-Spec, RT, SC, OOSA ERM, SA, SD V-Modell • • •Westmount I-CASE(Yourdon, OMT)
Westmount Techno-logy B. V. • VMS AD, DFD, ERD, CD, M-Spec, RT,
STDERM, OMT,SA, SD
ISOTEC, SSADM, V-Modell, (SETec) • • • • •
Legend:
• : available IE: Information Engineering PM: Project ManagementAD: Action Diagram CD: Class Diagram RT: Real-TimeBFD: Business Function Diagram CFD: Control Flow Diagram SA: Structured AnalysisDDD: Dialog Design Diagram CM: Configuration Management SC: Structure ChartsDFD: Data Flow Diagram C-Spec: Control-Specification SD: Struktured DesignELH: Entity Life History M-Spec: Mini-Specification SERM: Structured ERMERD: Entity-Relationship-Diagram MS: Module Structure SOM: Semantic Object ModelERM: Entity-Relationship-Modeling OAOD: OO Analysis/Design (Coad/Yourdon) STD: State Transition DiagramFT: Function Tree OOSA: OO System Analysis (Shlaer/Mellor) POM. Process Object ModellFctfld: Function Flow Diagramm OMT: Object Modelling Technique (Rumbaugh) VSOM: Vorgehensmodell für semantisches Objektmodell
Figure 1: Range of performance of the evaluated CASE-Tools (Part 2)
12 A Methodology of a CASE-Tool Assessment
© 1996 University of Cologne, Chair for Business Computing, Albertus-Magnus-Platz, D-50923 Köln, Germany
Criterium „Best Tool“ „Worst Tool“Product Name Annotation Product Name Annotation
User interface ProMod comfortable gra-phics editor
Teamwork lack of intuitivity
Development data storage ADW extensive Metamo-del
NEW(PREDICT CASE)
no online data-storage in networks
Team support Systems Engineer sophisticated me-chanisms for con-flict resolution
ObjectTool, OM-Tool, SOM
no mechanisms forconflict resolution
Openness / Expansibility Westmount Your-don
extensive possibili-ties for integrationof third-party pro-ducts
ObjectTool inadequate inter-faces to third-partyproducts
Adaptability Excelerator modifiable rulesand means ofdescription
case/4/0 mostly non-adaptable standardrules
Analysis Teamwork processing of tex-tual requirements
ObjectTool incomplete supportof analysis activi-ties
Design Systems Engineer powerful prototy-ping and databasedesign
Software throughPictures
incomplete supportof design activities
Realisation IEF generation ofexecutable systems
Software throughPictures, NEW(PREDICT CASE)
almost no generati-on
Documentation Innovator expressive reportgeneration functi-ons
ObjectTool many informationsonly available indiagrams
Project Management Maestro II completely integra-ted project mana-gement component
case/4/0, Object-Tool, OM-Tool,StP, SDW, SOM-CASE
almost no support
Quality Assurance Teamwork integrated testingcomponent
OM-Tool no or insufficienterror checks indiagrams, no testsupport
Method Support SOM Tool well-attunedto base method
ObjectTool incomplete avail-ability of means ofdescription
Uniformity Westmount OMT similar means ofdescription in allphases
TeamworkOOA/OOD
inconsistency bet-ween analysis anddesign
Function Range Maestro II support for allphases of the sy-stem’s lifecycle
OM-Tool mostly functionsfor drawing anddocumentation
Sophistication ADW continual impro-vement since 1989
OM-Tool UNIX-versiondiscontinued
Figure 2: Aspects of strength and weakness of the evaluated CASE-Tools
Figure 2 contains only relative statements. Concerning user-comfort, for example, no CASE-Tool
meets the level of standard software for graphic- and text-applications. Furthermore, the differen-
ces in performance are not consistent within the single categories. That leads to the fact, that the
tools are quite equal concerning the „classical“ CASE-duties analysis and design, while there is a
A Methodology of a CASE-Tool Assessment 13
© 1996 University of Cologne, Chair for Business Computing, Albertus-Magnus-Platz, D-50923 Köln, Germany
great deviation in efficiency concerning cross-section functions (projectmanagement, quality-
assurance etc.) and team support / interfaces.
Since a complete discussion of the survey findings regarding criteria and CASE-tools is not pos-
sible here, in the following chapter the CASE-tools will be grouped in order to describe the per-
formance of the examined products.
4.2 Tendency statements
Depending on the point of view, different tendencies become apparent regarding the examined
CASE-tools’ performance. In this scope, a suitable classification of the CASE-products is the
basis for meaningful tendency statements. In this survey, we use a formal classification and prefer
a two-dimensional grouping based on the study’s results.
On one hand, a classification is carried out according to the hardware and operating system plat-
forms:
PC-(LAN-)Tools: This includes all tools for (usually) networked IBM-compatible Personal Com-
puters using the operating systems DOS/Windows or OS/2.
• Workstation-Tools: This category includes tools for a UNIX-environment.
Tools for heterogenous hardware environments: This includes tools for a network with different
hardware and operating system platforms (e.g. Windows-PCs as front-ends and UNIX- or MVS-
hosts as back-ends).
On the other hand, tools are differentiated based on their underlying development methodology,
into tools for:
• conventional system development and
• object-oriented system development.
Object-oriented in the sense applied in this survey designates approaches which integrate the data-
and functional view, and in which an information system is composed of objects that communicate
via messages and can inherit certain properties from each other [for a more detailed distinction
between conventional and object-oriented systems development refer to e.g. Somm93].
Among the tools evaluated and grouped in this manner, it is possible to point out certain trends.
14 A Methodology of a CASE-Tool Assessment
© 1996 University of Cologne, Chair for Business Computing, Albertus-Magnus-Platz, D-50923 Köln, Germany
4.2.1 Tendency statements by hardware/OS environment
PC-(LAN-) Tools like case/4/0 have the advantage, among others, of a high degree of user-
friendliness at a comparatively low price (e.g. DM 799 for ProMod+). The number of available
CASE-Tools for the support of conventional methods on this development platform is very large.
Also, there is extensive support available for the earlier phases of systems development.
PC-(LAN-) Tools display weaknesses regarding response time (especially tools based on relatio-
nal databases) and regarding the available function range in later phases of systems development (
especially code generation, program test and maintenance). Additionally, none of the PC-(LAN)
Tools designated for object-oriented systems development can be regarded as suitable for practi-
cal application according to these criteria.
Tools for heterogenous environments like MAESTRO II and NEW/PREDICT-CASE, because
they allow simultaneous use of the potential of different hardware platforms and suggest themsel-
ves especially for systems development on mainframes by major users and offer the appropriate
function range.
However, the user must accept the disadvantage of a higher price as well as a higher system
complexity. Additionally, deficiencies concerning the data storage (e.g. lack of an online con-
nection of the PCs to the host) were found in the tools that aren’t designed as Client-Server-
Systems (e.g. NEW/PREDICT-CASE). Obviously, tools for heterogenous environments in most
cases allow only conventional systems development.
The Workstation-based CASE-Tools (like e.g. Westmount I-CASE) distinguish themselves by a
broad function range and a good selection of object-oriented methods. In addition, the response
time is excellent. The evaluation also showed that workstation-based tools (e.g. Teamwork or
Westmount OMT) are clearly superior to the PC-based ones when it comes to object-oriented
systems development.
Installation and operation of the workstation-based tools, however, created comparatively great
problems during the evaluation (e. g. Westmount) - which is in a major part due to the high de-
gree of system complexity. The prices are relatively high (the workstation-based Teamwork, for
example, costs about DM 60.000).
A Methodology of a CASE-Tool Assessment 15
© 1996 University of Cologne, Chair for Business Computing, Albertus-Magnus-Platz, D-50923 Köln, Germany
4.2.2 Tendency statements by development methodology
Typically, conventional CASE-Tools have been in the market for a substantially longer time than
the recently coming along tools supporting object-oriented system development. Along with this
fact comes a greater degree of sophistication in the market for conventional CASE-Tools which
becomes apparent in the following ways:
• Standard-Methods and means of description (ERM, SA/SD, etc.) which simplify the change
of tools because of less training expense and better possibilities for exchange between tools.
This also allows the combination of several tools.
• Stability and sophistication
• Broad function range
• Broad palette of development- and runtime-environments
Today, conventional tools are indeed less efficient concerning the completeness of the develop-
ment process, which is due to irregularities within the underlying methods.
The actual and predominant phenomenon observed at the market for object-oriented CASE-Tools
are the dynamics and immaturity of the market. In contrast to the conventional tools, there exists
no standard of methods at the moment. Three of the evaluated programs (Software Through
Pictures by IDE, Westmount OMT and OM-Tool by IQProducts) work according to the object
modelling techniques of Rumbaugh et al., one of them uses OOA/OOD by Coad/Yourdon
(ObjectTool, Rösch Consulting) and another one uses OOSA by Shlaer/Mellor (Teamwork, Ca-
dre). The immaturity becomes obvious with the short terms within the appearance of new relea-
ses. So it is no extraordinary thing, that for example Westmount OMT announced 3 new versions
in 1994.
This aspect shows, that manufacturers are still in the midst of their product development and the-
refore close to presenting the customer much better versions. Another aspect of immaturity is far
more unfavourable to the customer. It is not only that in rapid succession new manufacturers and
distributors are entering this market segment, it can neither be excluded, that the development or
distribution of a product will be stopped by manufacturers or distributers, because the product
16 A Methodology of a CASE-Tool Assessment
© 1996 University of Cologne, Chair for Business Computing, Albertus-Magnus-Platz, D-50923 Köln, Germany
had turned out to be immature and inefficient. For example, the distribution of OM-Tool for the
UNIX-platform had been stopped in the end of 1993. The further destiny of the Windows-version
is still unknown. Keeping in mind, which costs are caused by the selection and integration of
CASE-Tools, and what costs arise with the migration to other object-oriented tools after having
realized several projects, one can imagine, what influence a decision like that has on the user.
Thus, it is essential for the user, especially if more than only a pilot project shall be realized, to
examine the actual and future performance of the chosen manufacturer or distributer and his pro-
ducts very close.
There is still another reason for this critical investigation of products. The object-oriented CASE-
Tools have a noticably broader range of functionality and quality. For example, „Drawing-tools“
like ObjectTool and OM-Tool present only a few more functions, apart from drawing and docu-
mentation facilities, that are still partially immature. But now there are also programs available,
that are definately more efficient and have a higher performance. There are products that are still
in an experimental state, as well as already sophisticated tools. Thus, this situation is similar to the
situation with the conventional tools prevailing in the beginning of the 80´s. In the scope of this
study, no object-oriented tool could have been found, that covered the actual „fringe subjects“ as
project management, configuration management, etc. in a satisfactory way. Basically, no desired
function should be taken for granted when selecting and integrating object-oriented tools. Even
basical functions as team support are often not realized, as for example in ObjectTool, OM-Tool
and SOM.
Object-oriented methods claim to support reuse very well. In all the evaluated systems, this sup-
port is limited to inheritance. Exceeding reuse concepts that are surely indispensable to planned
and company-wide reuse, have not been discovered. There are no functions available that syste-
matically support the storage and tracing of classes or results of analysis and design.
The often mentioned completeness of the system development process indeed is an advantage of
the object-oriented system development. Iterative proceedings became more practicable and tools
ideally contribute one´s share, even though there are actually still missing some necessary functi-
ons for reverse engineering, such as conversion of code, e. g. from class libraries to diagrams.
Even today, the best object-oriented CASE-Tools present a very good connection to data bases
and C++-Compiler. The connection to interfaces though fails, due to the lack of standardization
and availability of object-oriented interface-tools. The reason is, that many tools of the data base
A Methodology of a CASE-Tool Assessment 17
© 1996 University of Cologne, Chair for Business Computing, Albertus-Magnus-Platz, D-50923 Köln, Germany
sector, that formerly have not been object-oriented, only now are being developed in the direction
of object-orientation.
Finally, it can be stated that since considerable time, there are tools available, that are entirely
suitable for the object-oriented development of pilot projects or single phase-specific resp. phase-
comprehensive activities (e.g. documentation). Before the respective implementation, one has to
be aware of the weak points of these tools and must check, whether they are acceptable. Soluti-
ons that are applicable during the whole developing procees and that present a complete replace-
ment for the conventional CASE-Environments, can only be found at the market in recent times.
4.3 Evaluation of the CASE-Tools using benefit-value-analysis
The final evaluation requires a weighting of the results of the corresponding criteria. This
weighting should be done in an individual way for each company. It may as well follow the
established quality standards as the ISO 9000 (standard for quality management systems). Multi-
plying the degree of fulfillment with the assigned weight of each criteria and then building a total
only allows to put the examined tools in an order of precedence. The following figure presents an
example of both, a weighting by a CASE-Tool user (a german insurance company1) and a
weighting derived from the software quality management system of ISO 9000 [see Herz94b].
1 The weighting has been fixed by Deutsche Krankenversicherung (DKV), Cologne, a german health insurancecompany.
18 A Methodology of a CASE-Tool Assessment
© 1996 University of Cologne, Chair for Business Computing, Albertus-Magnus-Platz, D-50923 Köln, Germany
Evaluation criteria Weighting done by a CASE-Tool-user of the insurance-sector
Weighting according to therequirements concerningsoftware development processof ISO 9000
Development data maintenance � � � � � � � � � � �meta-Model � � � � � � � � � Team support � � � � � � � � �
multi-user-ability � � � � � � � � conflict-solving-concept � � � � � � � � � usergroup-concept � � � � � � � � Interface
� � � � � � � � � � �
(standard-)interfaces � � � � � � � � adaptabilityt � � � � � � � � Analysis
� � � � � � � � � �
methods for analysis � � � � � � � textual requirements � � � � � � � � � Design
� � � � � � � � � �
methods for design � � � � � � � � � prototyping � � � � � � � � � reuse � � � � � � � � � completenesst � � � � � � � � � Realization
� � � � � � �
code-Generation � � � � � � � � � runtime-versions � � � � � � � � Documentation
� � � � � � � � � �
Quality assurance-manual � � � � � � � � � developer-documentation � � � � � � � � � user-documentation � � � � � � � � Quality-assurance � � � � � � � � � �examination / contract / project � � � � � � � � � check text + graphics � � � � � � � � testing � � � � � � � � Projectmanagement � � � � � � � � � � �
phases + activities � � � � � � � � � measurements � � � � � � � � Configuration management � � � � � � � � � � � �
management of versions � � � � � � � � � management of alternations and releases � � � � � � � � � status administration � � � � � � � � � retracability � � � � � � � � � Total � � � � � � � � � � �
Figure 3: Relative weighting of the evaluation criteria according to the requirements of an user of
the insurance sector, regarding ISO 9000
A Methodology of a CASE-Tool Assessment 19
© 1996 University of Cologne, Chair for Business Computing, Albertus-Magnus-Platz, D-50923 Köln, Germany
� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �
� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �
� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �
� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �
� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �
� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �� � �
� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �
� �� �� �� �� �� �� �� �� �� �� �� �� �� �� �
! !! !! !! !! !! !! !! !! !! !! !! !! !! !! !! !! !
" " "" " "" " "" " "" " "" " "" " "" " "" " "" " "" " "" " "" " "" " "" " "" " "" " "" " "" " "" " "" " "" " "" " "
# # ## # ## # ## # ## # ## # ## # ## # ## # ## # ## # ## # ## # ## # ## # ## # ## # ## # ## # ## # ## # ## # #
$ $$ $$ $$ $$ $$ $$ $$ $$ $$ $$ $$ $$ $$ $$ $$ $$ $$ $$ $$ $$ $$ $$ $$ $$ $$ $$ $$ $$ $$ $
% % %% % %% % %% % %% % %% % %% % %% % %% % %% % %% % %% % %% % %% % %
& & && & && & && & && & && & && & && & && & && & && & && & && & && & && & && & && & && & && & && & && & && & && & && & && & && & && & && & && & &
' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '' ' '
( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (( ( (
) )) )) )) )) )) )) )) )) )) )) )) )) )) )) )) )) )) )) )) )) )) )) )) )) )) )) )) )) )) )) )) )) )) )) )) )
* ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** *
+ + ++ + ++ + ++ + ++ + ++ + ++ + ++ + ++ + ++ + ++ + ++ + ++ + ++ + ++ + ++ + ++ + ++ + ++ + ++ + ++ + ++ + ++ + ++ + ++ + +
, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,, , ,
- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -
. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .
/ / // / // / // / // / // / // / // / // / // / // / // / // / // / // / // / // / // / // / // / // / // / // / // / // / // / // / // / // / // / // / // / // / // / // / // / // / // / /
0 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 0
1 1 11 1 11 1 11 1 11 1 11 1 11 1 11 1 11 1 11 1 11 1 11 1 11 1 11 1 11 1 11 1 11 1 1
2 2 22 2 22 2 22 2 22 2 22 2 22 2 22 2 22 2 22 2 22 2 22 2 22 2 22 2 22 2 2
3 3 33 3 33 3 33 3 33 3 33 3 33 3 33 3 33 3 33 3 33 3 33 3 33 3 33 3 33 3 33 3 33 3 33 3 33 3 33 3 3
4 4 44 4 44 4 44 4 44 4 44 4 44 4 44 4 44 4 44 4 44 4 44 4 44 4 44 4 44 4 44 4 44 4 44 4 44 4 44 4 44 4 44 4 44 4 44 4 44 4 44 4 44 4 44 4 44 4 44 4 4
5 55 55 55 55 55 55 55 55 55 55 55 55 55 55 55 55 55 55 55 55 55 55 55 5
6 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 66 6 6
7 7 77 7 77 7 77 7 77 7 77 7 77 7 77 7 77 7 77 7 77 7 77 7 77 7 77 7 77 7 77 7 77 7 77 7 77 7 77 7 7
8 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 88 8
9 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 99 9 9
: :: :: :: :: :: :: :: :: :: :: :: :: :: :: :: :: :: :: :: :: :: :: :: :: :: :: :: :: :: :: :: :: :: :: :: :: :: :: :
; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;; ; ;
44%
27%
42% 42%39%
53%
75%
22%
16%
26%
36% 35%
47%
21%
45%
62%
55%57%
63%
39%
56%
49% 49%
61%
79%
26%23%
31%
47%
37%
57%
31%
56%60% 61%
64%
0%
10%
20%
30%
40%
50%
60%
70%
80%
AD
W
case
/4/0
DD
B C
AS
E
Exc
eler
ator
II
IEF
Inno
vato
r
Mae
stro
Obj
ect T
ool
OM
-Too
l
NE
W (
PR
ED
ICT
CA
SE
)
Pro
Mod
-Plu
s
Sof
twar
e th
roug
h P
ictu
res
SD
W
SO
M-C
AS
E
Sys
tem
s E
ngin
eer
Tea
mw
ork
I-C
AS
E O
MT
I-C
AS
E Y
ourd
on
< << <ISO 9000= == =Insurance company
Figure 4: Fulfillment of requirements of the evaluated CASE-Tools according to an insurance
company as a user and the ISO 9000
Concerning the insurance company, the CASE-Tools of Softlab (MAESTRO II, 79%), West-
mount (I-CASE YOURDON, 64%) and Knowledge Ware (ADW, 63 %) are of special interest,
because they meet the requirements for the most part (see figure 4). Regarding ISO 9000, it is
completely different. The requirements for tool support of the software development process em-
phasize the point of supporting the documentation and quality assurance. But these demands are
only met by a few CASE-Tools (e. g. MAESTRO II), so that in the overall placing, the ADW
tool loses some of its importance (63 % to 44 %). This kind of approach illustrates the impact of
individually weighted evaluation criteria on the ranking of CASE-Tools, that have been taken into
account for an investment.
5 Conclusion
Concerning the validity of this examination, the continual improvement of CASE-Tools is rather
problematic. It led to the fact, that directly after the conclusion of this study in spring 1994, there
20 A Methodology of a CASE-Tool Assessment
© 1996 University of Cologne, Chair for Business Computing, Albertus-Magnus-Platz, D-50923 Köln, Germany
had been new releases of some of the evaluated tools. Furthermore, the definition of a tool as a
CASE-Tool becomes more and more difficult (e. g. due to the business process modeling). Final-
ly, subjective influences during the evaluation cannot entirely be excluded, in spite of all efforts to
gain objectivity.
The evaluation shows, that there are considerable differences within the various CASE-Tools that
indeed require a careful selection. The present data further allows broader examinations or bene-
fit-value-analysis concerning the question, which CASE-Tools meet the requirements of the V-
Model (standard processing model of the federal authority) or the ISO 9000 at best.
Independant of the environment and method of development, recent approaches concerning mo-
deling of tasks and business processes / business reengineering, have not yet become established
in the examined CASE-Tools [concerning software support of business reengineering, see
Spur94]. The authors´ opinion is, that the contentual and organizational aspects of business
process modeling, eventually using reference models, are an important factor of success for the
future of CASE. These aspects found consideration in the ARIS-Tools. This tool have not been
evaluated in this study. Manufacturers of CASE should therefore relate their product development
closer to the requirements of their customers, whose problems are rather located within the earlier
phases of the system development. They should not continuously try to expand the functionality
implementing the latest technological trends of design and realization [concerning customer-
oriented product development, see Akao92].
A Methodology of a CASE-Tool Assessment 21
© 1996 University of Cologne, Chair for Business Computing, Albertus-Magnus-Platz, D-50923 Köln, Germany
6 References
[Akao92]
Yoji Akao: Quality Function Deplyoment: Integrating Customer Requirements into Product De-
sign. Cambridge 1990
[ANSI90]
ANSI, IEEE (Hrsg.): ANSI, IEEE (Hrsg.): Standard Glossary of Software Engineering Termino-
logy. Std 610.12-1990. New York 1990
[Balz93]
Helmut Balzert (Hrsg.): CASE. Systeme und Werkzeuge. 5. Auflage, Mannheim u. a. 1993
[BeHe92]
D. Berkau, G. Herzwurm: Kriterien für die Auswahl PC-gestützter Software-Entwicklungsumge-
bungen - dargestellt am Beispiel von Excelerator, Information Engineering Workbench, ProKit
WORKBENCH und Systems Engineer. In: Information Management, Nr. 1, 1992, S. 42-55
[Gane90]
Chris Gane: Computer-Aided Software Engineering (The Methodologies, The Products, And The
Future). New Jersey 1990
[Harm94a]
Paul Harmon (Hrsg.): CASE and the move to objects. A strategic response to industry trends. In:
Application Development Strategies, Vol. 6, Nr. 3, 1993, S. 1-14
[Harm94b]
Paul Harmon (Hrsg.): The market for object-oriented software development products in North
America. In: Object-Oriented Strategies, Vol. 4, Nr. 4, 1993, S. 1-12
[Herz94a]
22 A Methodology of a CASE-Tool Assessment
© 1996 University of Cologne, Chair for Business Computing, Albertus-Magnus-Platz, D-50923 Köln, Germany
Georg Herzwurm (Hrsg.): CASE-Technologie in Deutschland. Orientierungshilfe und Marktüber-
blick für Anbieter und Anwender. Studien zur Systementwicklung des Lehrstuhls für Wirt-
schaftsinformatik der Universität zu Köln. Band 2. Köln 1994
[Herz94b]
Georg Herzwurm, Andreas Hierholzer, Michael Kunz: Eignung konventioneller und objekt-
orientierter CASE-Tools zum Aufbau eines Qualitätsmanagementsystems nach ISO 9000. In:
Information Management, Nr. 3, 1994, S. 72-76
[ISIE91]
ISO/IEC (Hrsg.): International Standard ISO/IEC 9126: Information Technology - Software Pro-
duct Evaluation. Quality Characteristics and Guidelines for their use. First Edition 1991-12-15,
Genf 1991
[Ovum93]
Ovum (Hrsg.): Case Products. London 1993
[Somm93]
Ian Sommerville: Software Engineering, 4. Edition, Wokingham 1992
[Spur94]
Kathy Spurr (Hrsg.): Software assistance for business re-engineering. Chichester, New York
1994.
A Methodology of a CASE-Tool Assessment 23
© 1996 University of Cologne, Chair for Business Computing, Albertus-Magnus-Platz, D-50923 Köln, Germany
Verzeichnis der Bände zur Reihe: Studien zur Systementwicklung
Herausgegeben von Prof. Dr. Werner Mellis, Köln; Dr. Georg Herzwurm, Köln; Dr. Dirk Stelzer,
Köln
Band 1: Georg Herzwurm, Uwe MüllerKriterienkatalog zur Unterstützung der Auswahl von CASE-Tools.Köln 1993
Band 2: Georg Herzwurm (Hrsg.)CASE-Technologie in Deutschland. Konventionelles und objektorientiertes CASE.Orientierungshilfe und Marktüberblick für Anbieter und Anwender.Köln 1994
Band 3: Werner Mellis, Georg Herzwurm (Hrsg.)CASE für Manager.Köln 1994
Band 4: Klaus SchmollingInnovative Methoden und Techniken bei der Entwicklung von Anwendungssoftware -Stand und Perspektive in deutschen Versicherungsunternehmen - Ergebnisse einer em-pirischen Untersuchung.Köln 1994
Band 5: Klaus SchmollingEmpirische Untersuchung zur Entwicklung von Anwendungssoftware in deutschenUnternehmen.Köln 1994
Band 6: Dirk Stelzer, Jan KoprowskiStand der Zertifizierung der 25 umsatzstärksten Softwareanbieter und DV-Berater nachDIN ISO 9001 in Deutschland - Ergebnisse einer Erhebung im Frühjahr 1994.Köln 1994
Band 7: Georg Bellin, Dirk StelzerSoftwarequalitätsmanagement gemäß ISO 9000. Ergebnisse einer empirischen Unter-suchung zertifizierter Qualitätsmanagementsysteme.Köln 1995
Band 8: Andreas Hierholzer, Georg HerzwurmMethodology of a CASE-Tool Assessment.Cologne 1996