crews.gif (2510 Byte)

Reports 1997
Home Up Reports 1999 Reports 1998 Reports 1997 Reports 1996 Registration

 

 

 

 

CREWS Reports 1997

 

S. Nurcan, C. Rolland

appeared in
Proceedings of the Workshop on the many facets of Process Engineering (MFPE’97), Gammarth, Tunis, September 22-23, 1997.

Abstract
The Computer Supported Cooperative Work (CSCW) discipline examines the possibilities and effects of technological support for humans involved in collaborative group communication and work processes. Organizations are built on the principle that groups of people can carry out tasks which are not feasible individually. For many of them, well-structured and ill-structured procedures coexist in work processes and must be managed in the final solution. Cooperative work techniques become very important in organizations. One can note the emergence of many products dedicated to developing cooperative systems. Therefore, it is necessary to emphasize the specificities of these applications in order to take them into account as soon as possible during design. This paper presents a process meta-model for goal-driven cooperative work processes.

 

Klaus Weidenhaupt, Klaus Pohl, Matthias Jarke, Peter Haumer, CREWS Team

appeared in:
IEEE Software, March, 1998 and
International Requirements Engineering Conference (ICRE’98), Colorado Springs, Colorado, USA, April 6-10, 1998.

Abstract
Scenario-based approaches are becoming ubiquitous in systems analysis and design but their definition and scope remain vague. Complementing the recently proposed CREWS classification framework for scenario-based RE, this paper reports on an exploratory survey of practice conducted through site visits with 15 projects in four European countries. The main findings include that: (1) the variety of purposes and uses of scenarios in the process is much greater than expected; (2) as a consequence, we must take scenarios much more seriously as important design artifacts, offering better means for structuring, management, and evolution. To handle these complex processes, users request more explicit methodological guidance and more adequate tool support.

 

Camille Ben Achour, Colette Rolland

submitted to:
International Requirements Engineering Conference (ICRE’98), Colorado Springs, Colorado, USA, April 6-10, 1998.

Abstract
This paper attempts to define the role and the constraints of Natural Language (NL) interpretation within the Requirement Engineering (RE) process. In so doing, it addresses two key features of a NL interpretation tool. First, genericity towards the design modelling language targeted by the interpretation can be supported owing to the semantic level of analysis of the input scenario. Genericity allows to interpret any kind of textual scenario into any kind of target design model. The second characteristic of the interpretation tool is its modularity which permits to automate partly RE tasks. Tightly embedded in the RE process, the interpretation becomes a powerful support that participates efficiently to the automation of numerous RE tasks. We illustrate this feature with an example making use of the interpretation tool for the improvement of the consistency between sequence diagrams and textual scenarios.

 

N.A.M. Maiden, S. Minocha, K. Manning, M. Ryan

appeared in:
International Requirements Engineering Conference (ICRE’98), Colorado Springs, Colorado, USA, April 6-10, 1998.

Abstract
CREWS-SAVRE is a first-draft method and software tool for systematic scenario generation and use. This paper reports on two interleaved strands of research and development. The first involves theoretical research into classes of exceptions in software-intensive systems. The second is the development of a software prototype which has been used to acquire requirements from current scenario users. These user requirements are presented and implications for the second version of the prototype are discussed. The paper ends with advantages of integrating basic software engineering research with user-centred system design.

 

Patrick Heymans

Abstract
Writing requirements specifications of complex composite systems makes it necessary to have a language which is both formal to allow reasoning on specifications and declarative enough to allow the analyst to express himself in a natural way. Albert II is a language that tries to achieve these goals and, besides that, provides templates that guide the analyst in writing the specification. However, because of its formality, the resulting specification cannot in general be read by the various stakeholders. Validation tools are therefore required, among which an animator. In this paper we aim at giving an overview of the functionalities and the architecture of the animator for Albert II specifications.

 

Klaus Pohl, Klaus Weidenhaupt

appeared in
Proceedings of the 6th European Software Engineering Conference / 5th ACM SIGSOFT Symposium on the Foundations of Software Engineering (ESEC/FSE 97), Zurich, September 22-25, 1997.

Abstract
Research in process-centered environments (PCEs) has focused on project management support and has been dominated by the search for suitable process modelling languages and enactment mechanisms. The consequences of the process orientation on the tools used during process performance, and for offering fine-grained, method-based support to the engineers performing the process have been studied much less. In this paper, we discuss the requirements for a tighter integration of interactive engineering tools and present a contextual approach for the process-integration of those tools. To achieve process integration we argue that tools, like processes, should be explicitly defined. The integration of the tool models with the process definitions forms an environment model which is interpreted during tool execution. Based on this interpretation tool behavior is adjusted according to the process definition; i.e. the interpretation empowers the tools to provide fine-grained method-conform process support. Our approach has been implemented as a reusable object-oriented framework and validated by specializing this framework to develop two prototypical process-integrated environments (PIEs).

 

R. Motschnig-Pitrig, H.W. Nissen, M. Jarke

appeared in
Proc. of the 9th Intl. Conf. On Software Engineering and Knowledge Engineering (SEKE ’97), Madrid, Spain, June 17-20, 1997.

Abstract
Semi-formal methods for requirements and design suffer from a problem of scalability. They offer different notations to describe complementary abstractions of the problem to be analized, but each such abstraction in itself becomes rapidly much too big to be handled in one piece. Some methods therefore offer basic view mechanisms, variously called modules, categories, viewpoints etc., to introduce smaller units. However, they lack in guidance what views to define and how to interrelate them in a systematic manner. The problem we address in this paper can be simplistically stated as: "Given that a multi-notation specification must be composed from hundreds of fragments each no larger than one page, according to what facets should we acquire, classify, compose/decompose, and compare these views?" We develop a situational framework of such facets, and formalize it in a metamodel. We compare three basic process models of how the framework can be applied, and describe the formal and technical support required for each of these models.

 

Matthias Jarke, Klaus Pohl, Peter Haumer, Klaus Weidenhaupt, Eric Dubois, Patrick Heymans, Colette Rolland, Camille Ben Achour, Corinne Cauvet, Jolita Ralyté, Alistair Sutcliffe, Neil Maiden, Shailey Minocha

Abstract
This paper reports on the results of 15 site visits which aimed at gaining explorative insights into scenario use in current industrial practice. Moreover, highlights of an follow-up questionnaire disseminated to a larger number of organizations are presented.

 

Camille Ben Achour

appeared in
Proceedings of the Third International Workshop on Requirements Engineering: Foundation for Software Quality (REFSQ'97), Barcelona, Catalonia, June 16-17, 1997

Abstract
Scenario based approaches intend to include end stakeholders in the Requirements Engineering process by the use of adapted expression media such as natural language. However if they describe how to use scenarios for the elicitation, exploration, or validation of requirements, very few approaches say how scenario could be integrated into different existing methods and supported by tools. Based on a survey on linguistic based requirements engineering instruments, this paper proposes to adapt two linguistic theories, namely Grice's dialogue theory and on Fillmore's Case Grammar, respectively to support the dialogue on scenario, and to automate the interpretation of scenario described in natural language. Described as generic chunks of processes, these instruments are intended to be adapted to any method and to any scenario based approach of Requirement Engineering relying on the use of text.

 

Jolita Ralyte and Camille Ben Achour

appeared in
Proceedings of the Workshop on the Many Facets of Process Engineering (MFPE'97), Gammarth, Tunis, September 22-23, 1997.

Abstract
Scenario Based Requirement Engineering (RE) techniques have for main motivation the opening of Requirement Engineering processes to neophytes. By combining the methodological approach of requirement analysis to informal descriptions, they increase productivity into classical RE techniques. In this context, an important problem is the integration of scenario approaches into RE methods. This paper presents a knowledge base of different scenario chunks which allows to identify the scenarios appropriate to the context of the RE project, and to integrate them into any RE methods. The identification of scenario chunks is partly based on the internal properties of scenarios which are extensively described in a framework for scenario classification.

 

Selmin Nurcan and Colette Rolland

appeared in
Proceedings of the Seventh European-Japanese Conference on Information Modeling and Knowledge Bases, Toulouse, France, May 27-30, 1997

Abstract
Cooperative work techniques are becoming very important in organisations as well as in the information systems community. The Computer Supported Cooperative Work(CSCW) discipline makes the assumption that collaborative work and processes can be supported by software tools. This requires among others to develop models able to represent cooperative work processes. In this paper we propose a meta-modeling framework to deal with a range variety of CSCW models. We present and exemplify a meta-model from which models can be instantiated. The meta-model is taylored to support the modeling of both well-structured and ill-defined work procedures and their interactions.

 

Colette Rolland

appeared in
Proceedings of the INFORSID Conference (INFormatique des ORganisations et Systemes d'Information et de Decision) Toulouse, France, June 10-13, 1997

Abstract
The area of method engineering has emerged in response to an increasing feeling that methods are not well suited to the needs of their users, the application engineers. Method engineering aims at developing methods. The paper uses the four worlds framework to investigate the subject of method engineering. It presents a state-of-art survey with a view to identifying the important problems and research directions being followed to solve these. The major problems not yet addressed or which have newly emerged are highlighted.

 

Klaus Pohl and Peter Haumer

appeared in
'Third International Workshop on Requirements Engineering: Foundation for Software Quality RESFQ', June 16-17, 1997, Barcelona, Spain

Abstract
Scenario-based approaches have proven useful for requirements elicitation, validation and negotiation. Besides the direct and indirect stated requirements current scenario-based approaches capture also contextual information about the existing or future system, but lack in a systematic support for representing and reasoning about this information.
Based on a literature survey we define a comprehensive set of concepts needed to represent contextual usage knowledge of scenarios. In contrast to existing approaches, we propose to relate contextual knowledge not only to the whole scenario, but also to the scenario components, e.g. single interactions between the system and the user of the system. Consequently, we propose two contextual models, a scenario context model (SCM) and an interaction context model (ICM).

 

Patrick Heymans

appeared in
Proceedings of the Doctoral Consortium of the third IEEE International Symposium on Requirements Engineering (RE'97), Annapolis, MD, USA, January 6-10, 1997

Abstract
Formally and declaratively specifying requirements on real-time composite systems requires validation by stakeholders. One way to perform such validation is allowing people to experience the dynamic properties of the system to be built by using an animation tool. In this paper, we give an overview of the issues raised by the development of such a tool for the ALBERT language with the main goal of obtaining feedback on research that is a its very early stages.

 

Klaus Pohl, Ralf Dömges and Matthias Jarke

appeared in
Proceedings of the 9th International Conference on Advanced Information System Engineering, Barcelona, Spain, June 18-20, 1997.

Abstract
Traceability is a prerequisite for managing the evolution of (software) systems. Assuring overall traceability of a system development process, i.e., capturing and interrelating all possible data, is almost impossible and by far too expensive and labor intensive. To minimize the information to be recorded and to reduce the additional costs the types of trace information to be captured should be adjusted to project-specific needs, e.g., intended trace usage, time and money available.
In this paper we present an approach which supports method-driven trace capture. The project manager defines the trace information and the trace steps required for recording this information according to the actual needs in explicit traceability (method) models. In addition, the trace steps are integrated with the method definition used to guide the product development process. Based on the so gained extended method definition, the stakeholders are guided in capturing the defined trace information and trace capture is even partially automated. We report on experiments made with the prototypical implementation of the approach and discuss possible extensions.

 

Neil Maiden, Shailey Minocha, Keith Manning and Michele Ryan

appeared in
'Third International Workshop on Requirements Engineering: Foundation for Software Quality RESFQ', June 16-17, 1997, Barcelona, Spain

Abstract
Scenarios, in most situations, are descriptions of required interactions between a desired system and its environment which detail normative system behaviour. There is considerable current interest in the use of scenarios for acquisition, elaboration and validation of system requirements. However, despite this interest, there remains a lack of methods and software tools to generate and use scenarios during the requirements analysis phase. In this paper, we outline the architecture of a toolkit for semi-automatic generation of scenarios. We have derived complex taxonomies of exceptions to help a requirements engineer to predict non-normative system behaviour in a scenario. We have outlined a method of cause-consequence analysis to explore the occurrence of problem exceptions and their effects on system behaviour.

 

Neil Maiden, Shailey Minocha, Michele Ryan, Keith Hutchings and Keith Manning

appearded in 'Proceedings of the Workshop on Human Error and Systems Development', 19-22nd March, 1997, Glasgow University, Scotland, United Kingdom, organised by Dr. Chris Johnson

Abstract
Scenario based requirements analysis is an inquiry based collaborative process which enables requirements engineers and other stakeholders to acquire, elaborate and validate system requirements. A scenario, in most situations, describes the normative or expected system behaviour during the interactions between the proposed system and its environment. To account for non-normative or undesired system behaviour, it is vital to predict and explore the existence or occurrence of 'exceptions' in a scenario. Identification of exceptions and inclusion of additional requirements to prevent their occurrence or mitigate their effects yields robust and fault-tolerant design solutions.
In this paper, we outline the architecture of a toolkit for semi-automatic generation of scenarios. The toolkit is co-operative in the sense that it aids a requirements engineer in systematic generation and use of scenarios. The toolkit provides domain knowledge during requirements acquisition and validation of normative system behaviour. It also provides systematic guidance to the requirements engineer to scope the contents of a scenario.
Furthermore, we have identified three kinds of exceptions: generic, permutation and problem exceptions, and have derived complex taxonomies of problem exceptions. We propose to populate the toolkit with lists of meaningful and relevant 'what-if' questions corresponding to the taxonomies of generic, permutation and problem exceptions. The exceptions can be chosen by the requirements engineer to include them in the generated scenarios to explore the correctness and completeness of requirements. In addition, the taxonomies of problem exceptions can also serve as checklists and help a requirements engineer to predict non-normative system behaviour in a scenario.

 

 


Reports 1999 Reports 1998 Reports 1997 Reports 1996 Registration