🚀 go-pugleaf

The proceedings of the Fourth Workshop on the Next Generation of
CASE Tools (NGCT'93) are available as a technical report from the
Center for Telematics and Information Technology, University of
Twente.

Price: Nfl 45, US$ 25 (including shipping and money transfer)

Order by sending a message including a POSTAL ADDRESS to:
Sjaak Brinkkemper
CTIT
E-mail: sjbr@cs.utwente.nl

*******************************************************
*      Proceedings of the Fourth Workshop on the      *
*           Next Generation of CASE Tools             *
*     Universite Paris 1 Sorbonne - 7/8 June 1993     *
*******************************************************

Editors: S. Brinkkemper and F. Harmsen
Center for Telematics and Information Technology
University of Twente
the Netherlands
174 pages


Abstract

The Workshop on the Next Generation of CASE Tools (NGCT) is an
annual event, bringing together leading researchers on Computer
Aided Software Engineering (CASE). NGCT workshop is a pre-conference
workshop of the annual Conference on Advanced Information Systems
Engineering (CAiSE). The goal of this year's workshop, held in
Paris, is to conduct an in-depth discussion of research approaches
in the area of Computer Aided Software Engineering. Three main
themes have been identified: 
*       CASE architectures
*       Development process support
*       Advanced requirements engineering
The workshop committee accepted fourteen papers, which are grouped
in the proceedings according to these three themes. Among the topics
of the papers are: multiparadigm specification for interoperable
information systems, capturing design decisions, automated user
interface derivation, deductive repositories, human error analysis,
and business modelling.








ABSTRACTS OF PAPERS
---------------------

A Multiparadigm Environment for the Design of Interoperable
  Information Systems
I. Petrounias and B. Theodoulidis
Department of Computation - UMIST - Manchester - UK

Over the last twenty years, a large number of information systems 
development approaches incorporating sophisticated real-world 
representation mechanisms, have been developed. Despite the 
extensive use of advanced modelling orientations however, not a 
single paradigm has been proved to be sufficient for all the 
possible complexities of some application domain. This paper 
introduces the MEDIS multiparadigm environment for the design of 
interoperable information systems development environment. In 
general, a multiparadigm development environment would allow the 
concurrent use of different paradigms for specifying and designing 
the various aspects of the information system and at the same time 
it could guarantee the interoperability of the different models. The 
paper discusses the components of the MEDIS modelling formalism and 
the representation of the data modelling component into structures 
of the MEGALOG+ multiparadigm database platform.


Towards an Integration of Evolving Information Systems and 
  CASE-Tools
H.A. Proper
Department of Information Systems - University of Nijmegen - the 
Netherlands

In this article, we link CASE-Tool technology to the research 
concerning evolving information systems. First, an introduction to 
the notion of evolving information system is provided. Then, we 
relate CASE-Tools and evolving information systems to each other 
from two different points of view, by looking how each one can serve 
the other. 


Automated support for Conceptual to External Mapping
L. Campbell and T. Halpin
Department of Computer Science - University of Queensland - 
Australia

It is widely recognised that the process of mapping conceptual 
objects into the internal level can (for the most part) be 
automated. The process of using a conceptual schema to automatically 
generate a prototype at the external level (the user interface) has 
been given comparatively little attention. Various CASE tools do 
provide primitive generation of form templates from the logical 
schema (e.g. Oracle's CASE Designer). This paper describes a CASE 
procedure for generating application forms from the conceptual 
schema which is more complete in terms of the constraints 
considered, and which uses repository links to maintain consistency 
between the three architectural levels.
 

An Object-Oriented Approach for the Development of CASE tools
G.M. Wijers and E.R. Nieuwland
Software Engineering Research Centre - Utrecht - the Netherlands

In this article we present an approach to the problem of flexible 
and adaptable CASE tools which tries to get away from a given 
distinction between generic and specific functionality by offering a 
smooth transition from generic to specific functionality. We present 
an object-oriented approach that is used to specify the required 
functionality of a CASE environment. High-level object class 
definitions describe generic functionality. By means of inheritance 
specific functionality can be defined reusing more generic object 
class definitions. If necessary, given functionality can be 
overruled. 
We start the article with a definition of our approach. Although we 
have the feeling that the approach is also useful in other areas 
than CASE, it has been particularly developed for the construction 
of CASE tools. In the paper we also present an elaborated example of 
tools that have been built using the approach developed.
** paper not in the proceedings; available from the authors 
at gwijers@serc.nl **

Capturing Design Intent in Group Software Development
D. Jenkins
Department of Computer Science - University of Paisley - UK
** no abstract available **


Process Models for CASE Shell Environments
T. Araujo and M. Rossi
INESC - Lisboa - Portugal  and  University of Jyvaeskylae - Finland

This paper address some issues that are raised by the advent of CASE 
shells. The CASE shells have developed into a state, where the 
systems can provide support various description languages, but they 
have no, or very limited capabilities for process support, i.e. 
supporting the user in his/her tasks during the information systems 
development.
We develop here ideas for meta-activity models, i.e. models of the 
process parts of the software development methods, in a CASE shell 
environment. As an example a meta-data model for data-flow diagrams 
is developed, we then apply process models on this meta-data model 
and then provide some suggestions on its uses. In the discussion 
section some possibilities for the use of the developed models are 
represented. 


Deductive Repository Technology: Supporting Effective Domain
  Modelling and Validation
B. Theodoulidis  and P. Alexakis
Department of Computation - UMIST - Manchester - UK

This paper argues that substantial benefits can be accrued from the 
use of a deductive repository in the next generation of CASE Tools. 
To this end, the paper discusses a knowledge representation 
formalism for the capturing and representation of domain models 
incorporating aspects such as time modelling and complex objects 
that deals with three levels of abstraction namely, model meta-
knowledge, application knowledge and extensional knowledge.
The paper proposes a representation mechanism of the externally 
(user-oriented) representation of domain knowledge onto a deductive 
specification and demonstrates that validating the captured 
knowledge can be greatly facilitated through the use of a deductive 
repository.
 

Computer Aided Method Engineering based on Existing Meta-CASE
  Technology
F. Harmsen and S. Brinkkemper
CTIT - University of Twente - the Netherlands

This paper deals with various aspects of automated support for 
Method Engineering, established by so-called Computer Aided Method 
Engineering (CAME) tools. Using a CAME tool, Method Engineering is 
supported with respect to the storage, retrieval and assembly of 
method fragments. Fragments of several existing information systems 
development methods, that have been stored in a so-called method 
base, are selected by the tool using characterisation rules. The 
selected fragments are assembled according to certain consistency 
rules, which are also offered by the tool. After an introduction to 
Method Engineering, and in particular to method assembly, 
requirements of a CAME tool are described. In order to be able to 
meet these requirements, an architecture is described that considers 
the CAME tool to be an extension of a meta-CASE tool. Furthermore, 
important research issues concerning CAME are identified.


The Modelling of Human Error Generation Processes in Software               

Development
S. Komiya and Y. Ishizaki
Software Technology Center - IPA - Tokyo - Japan

There is a strong need to maintain a high level of reliability in 
software. The essential factor behind reliability problems in 
complete software products is that software development involves the 
efforts of many people. The software engineering approach to this 
problem is to eliminate human errors by mechanizing many parts of 
the software development process. However, there are still important 
parts of the development process that can only be accomplished by 
humans. Thus there is a need to determine and control the human 
factors that contribute to the loss of reliability in software 
development processes that are particularly difficult to automate. 
Therefore, authors intend to determine those human factors by 
experiment in order to eliminate the human factors that lower 
software reliability. This paper describes a model for human factors 
that affect software reliability and discusses what kind of methods 
can be used to verify such a model so as to establish a premise for 
experiments in software development. 


Automated Support for Verbalization of Conceptual Schemas
T.A. Halpin and J. Harding
Department of Computer Science - University of Queensland - 
Australia  and  ServerWare - Washington DC - USA

Most CASE tools for information systems modeling allow conceptual 
schemas to be largely specified in a graphical language. With 
varying degrees of automation the conceptual schema may then be 
mapped to internal and external levels. Some workbenches also 
support a textual conceptual schema language which might be used to: 
verbalize the graphical constructs; specify additional constraints 
and rules; query the conceptual schema directly. If such a language 
is to be of real assistance in the modeling process, it needs to be 
formal, expressive and natural. The FORML language was designed with 
these criteria in mind, and its textual version appears to have 
comparative advantages, as far as static data modeling is concerned. 
It features ordered, mixfix predicates of arbitrary arity with 
structured object terms, carefully chosen keywords to enable 
constraints and rules to be expressed naturally yet rigorously, and 
is currently supported in the InfoViews workbench.


Semantic Case Analysis of Informal Requirements
B. Belkhouche and J. Kozma
Computer Science Department - Tulane University - New Orleans - USA

Semantic case analysis provides a useful and powerful tool for an 
object-oriented analysis of software requirements. Two important 
areas of object-oriented requirements analysis are addressed: (1) 
identification of entities which should be modeled as objects in the 
software design; and (2) detection of ambiguity, inconsistency and 
incompleteness in the requirements documents. Available heuristics 
to identify these entities are based on coarse criteria, relegating 
much of the endeavour to intuition. We propose a set of refined and 
systematic heuristic rules to enhance the identification procedures. 
The proposed extensions further improve the requirements analysis 
process by focusing on basic verbs which are relevant to software 
design. The implemented system is capable of parsing the input 
requirements and extracting the objects, actions, and attributes for 
use in creating an object-oriented model of the requirements. A 
question-answering capability is provided to allow the user to ask 
specific questions about the model extracted from the requirements. 
Serving as a mediator, this system can lessen the chance for 
miscommunication by pointing out potential ambiguities, 
inconsistencies, imprecision, and other problems that can lead to 
misunderstanding between the client and the analyst.


Software Process Management System Supporting the Cooperation
  between managers and developers
A. Hazeyama and S. Komiya
Software Technology Center - IPA - Tokyo - Japan

The authors aim at constructing an integrated software project 
management system with an object database. This paper describes a 
model for project software management that forms the basis of the 
system. In accordance with the plan that have been made based on 
this model, this paper describes a framework to induce a series of 
tasks that a project manager has to perform as follows: giving work 
instructions based on the plan to workers, progress ascertainment 
based on progress reports sent by workers, analysis of the impacts 
by problems detected. In order to realize these facilities, the 
authors propose the progress of a single activity using the 
attribute "Status" and its value. And the authors show the 
transition of it. The status is not simply used as an indicator of 
the progress of an activity. Notification facilities are also 
realized which notify appropriate activities and persons of 
necessary information at the necessary time when the actions which 
change the status value of an activity are triggered, by referring 
to an activity network (which is the instantiation of the process 
model). For example, when a project manager gives work instructions 
to workers based on the work plan, the system decides whether 
activities are able to be performed or not on the schedule 
restrictions that the workers have. If the activities cannot be 
performed, the system notifies the situations. Using this mechanism, 
the authors believe it is possible to effectively support a manager 
of the project in which multiple activities are performed in 
parallel.


Business Modelling and Rule Capture in a CASE Environment
B. Wangler, R. Wohed and S. Oehlund
SISU - Kista - Sweden

This paper addresses the problem of business rule capture in IS-
development. In particular it discusses the role played by business 
modelling and goal analysis as a starting point for rule 
acquisition. As a background it starts by discussing business rules 
and business modelling from a general point of view. It concludes by 
presenting three pragmatic techniques for rule elicitation.