. Requirements engineering (RE) is perceived as an area of growing importance. Due to the increasing effort spent for research in this area many contributions to solve different problems within RE exist. The purpose of this paper is to identify the main goals to be reached during the requirements engineering process in order to develop a framework for RE. This framework consists of the three dimensions: . the specification dimension . the representation dimension . the agreement dimension Looking at the RE research using this framework, the different approaches can be classified and therefore their interrelationships become much clearer. Additionally the framework offers a first step towards a common understanding of RE. + This work was supported by ESPRIT Basic Research Action 6353 (NATURE) which is concerned with Novel Approaches to Theories Underlying Requirements Engineering and by the state Nordrhein-Westfalen, Germany. 1 Introduction There is general agreement among softwa...

ion-based Analogical Inference, in Analogical Reasoning, edited by D.H. Helman, Kluwer Academic Publishers,p. 147-170 [Grosz & Rolland 1990]: G. Grosz, C. Rolland; Using artificial intelligence techniques to formalize the information system design process; Proc. Intl. Conf. Databases and Expert Systems Applications, 374-380. [Grosz & Rolland 1991] G. Grosz, C. Rolland; Why and how should we hide conceptual models; Proc. 3rd Intl. Conf. Software Engineering and Knowledge Engineering (SEKE), Skokie, USA, 28-33 [Guindon 1990] Guindon R.: Designing the design process: exploiting opportunistic thoughts, Human-Computer Interaction Journal 5, 305-344 [Guindon & Curtis 1988] Guindon R. and Curtis B.: Control of cognitive processes during software design: what tools are needed? Proc. ACM-CHI'88, 263-269 [Hagelstein 1988] Hagelstein J.: Declarative approach to information systems requirements. Knowledge Based Systems, 1, 4, 211-220 [Hahn et al. 1991] U. Hahn, M. Jarke, T. Rose: Teamwork support ...

Abstract. This paper defines and analyses a computational model of similarity which detects analogies between objects based on conceptual descriptions of them, constructed from classification, generalization relations and attributes. Analogies are detected(elaborated) by functions which measure conceptual distances between objects with respect to these semantic modelling abstractions. The model is domain independent and operational upon objects described in non uniform ways. It doesn’t require any special forms of knowledge for identifying analogies and distinguishes the importance of distinct object elements. Also, it has a polynomial complexity. Due to these characteristics, it may be used in complex tasks involving intra or inter-domain analogical reasoning. So far the similarity model has been applied in the domain of software engineering. First, to support the specification of software requirements by analogical reuse and second, to enable the integration of requirements specifications, generated by the multiple agents involved in information system development. Details of these applications can be found in sited references. Also, we have conducted an empirical evaluation of: (i) the consistency of the estimates generated by the model against human intuition about similarity and (ii) its recall performance in tasks of analogi-cal retrieval, the results of which are presented in this paper. 1.

Reuse of software components is a must to enhance productivity in software engineering. Object-oriented techniques help to produce reuseable components. However, the problem of retrieval from a repository of reuseable components is difficult. CASE tools of today aid in this task, but are not sophisticated enough. Reuse of components by analogy seems to be an important element in the experienced software engineer's problem solving strategy. ROSA is a project that aims to use techniques from artificial intelligence, machine learning, and in particular analogical reasoning to help retrieving solutions to old software engineering projects for reuse in new projects. This report sets the context of the project and defines its content and boundaries.

Reuse, system integration, and interoperability create a growing need for capturing, representing, and using application-level information about software-intensive systems and their evolution. In ESPRIT Basic Research Project NATURE, we are developing an integrative approach to requirements management based on a threedimensional framework which addresses formalism as well as cognitive and social aspects. This leads to a new requirements process model which integrates human freedoms through allowing relatively free decisions in given situations. Classes of situations and decisions are defined with respect to the three-dimensional framework through the integration of informal and formal representations, theories of domain modeling, and the explicit consideration of nonfunctional requirements in teamwork. Technical support is provided by a conceptual modeling environment with knowledge acquisition through interactive as well as reverse modeling, and with similarity-based querying. 1

There is evidence that analogical reasoning play an important role in problem solving within software engineering. When we use analogical reasoning in the early phases of software development, we can reuse knowledge from a known analogous case to support both the modeling and implementation phases of a software development process. In order to support this in a tool, we need a mechanism for establishing analogies between a present case and previous cases. The first stage in analogy matching is the base filtering process. This process selects relevant cases to be tried for match. Indexing of cases is the standard approach and is here done along dimensions each describing properties of models or model components. These properties are identified using heuristics applied in the matching phase of the analogy search. Related values for a property in two cases will indicate support for a match on the basis of a particular heuristic. The most important heuristic used in analogical matching com...

Most software reuse research has ignored the role of the software engineer. However, software engineers tend to be better reasoners and have more experiences to recall than tool-based reuse mechanisms. This paper argues for integrating software engineers into existing reuse paradigms and providing tool support to assist problem description and component understanding, selection and adaptation. However, empirical studies indicate that these reuse tasks are difficult, even for experienced software engineers. Therefore, guidelines and a high-level architecture for design of tool support are based on reports of behaviour and problems arising during reuse. 1: Introduction Many technical solutions to software reuse problems have failed to result in widespread reuse. One reason may be that these solutions ignore human involvement. This paper investigates specific problems which arise from excluding software developers from the reuse process and proposes solutions based on human involvement fo...

ions in Requirements Engineering: an Exemplar Approach N.A.M. Maiden & A.G. Sutcliffe Department of Business Computing School of Informatics City University London Abstract This paper reports an intelligent advisor which assists software engineers to reuse domain abstractions to improve the consistency, completeness and clarity of requirement specifications. Understanding unfamiliar domain abstractions can be difficult, so partial exposure and visualisation of concrete examples and metaphors are proposed to aid comprehension prior to reuse. These strategies are incorporated into an iterative fact acquisition and domain retrieval dialogue with important implications for fact capture and modelling during requirements engineering. The effectiveness of this paradigm is shown during user studies with a prototype of the intelligent advisor, during which software engineering retrieved and understood correct domain abstractions whilst analysing a new domain. 1: Introduction Requirements engine...

The ROSA project aims to support reuse of object-oriented analysis models by analogy. An object can play a set of roles in a system. The identification of roles and their responsibilities seem to be useful for analogical reuse. For modelling the OOram method has been chosen due to its emphasis of roles and their behaviour. Through a discussion of two examples, we identify model components that can be reused, and a faceted classification of these components is proposed. The component categories are full models, sub-models, scenarios, messages, and roles. These categories are assigned a set of facets, where the two most important for reuse of analysis models are believed to be the component's name and the stimulus role within a role model; i.e the role that initiates some activity in a model. Keywords: Software reuse, Object-orientation, Artificial Intelligence, Analogy, Role modelling Repository Report Generator Reuse Facility Modelling Tool Code Generator Text Search Browser Analogy ...

Domain abstraction reuse during requirements engineering was evaluated as a paradigm during use of a research tool. Iterative fact acquisition and abstraction retrieval, supported by example-based explanation of abstractions, was implemented in a prototype known as AIR. Observation of AIR's use by inexperienced software engineers revealed this paradigm's effectiveness despite several problems arising. These problems were examined to identify improvements to future versions of intelligent requirements engineering environments. 1. Introduction Requirements engineering is complex, error-prone and in need of intelligent tool support (e.g. [1]) to assist the capture, modelling and validation of requirements. One solution is to populate tools with generic domain abstractions to assist requirements specification [2] [3] [4]. This paper proposes a user-centred paradigm for reusing such abstractions and a tool called AIR (Advisor for Intelligent Reuse) to implement the paradigm. This paper prop...