This test is a pre-print of Visser, W. (1992). Designers ’ activities examined at three levels: organization, strategies & problem-solving. Knowledge-Based Systems, 5(1), 92-104.

This paper presents a state-of-art review of empirical research on object-oriented (OO) design. Many claims about the cognitive benefits of the OO paradigm have been made by its advocates. These claims concern the ease of designing and reusing software at the individual level as well as the benefits of this paradigm at the team level. Since these claims are cognitive in nature, its seems important to assess them empirically. After a brief presentation of the main concepts of the OO paradigm, the claims about the superiority of OO design are outlined.

This article presents a cognitively oriented viewpoint on design. It focuses on cognitive, dynamic aspects of real design, i.e., the actual cognitive activity implemented by designers during their work on professional design projects. Rather than conceiving designing as problem solving—Simon’s symbolic information processing (SIP) ap-proach—or as a reflective practice or some other form of situated activity—the situativity (SIT) approach—we consider that, from a cognitive viewpoint, designing is most appropriately characterised as a construction of representations. After a critical discussion of the SIP and SIT approaches to design, we present our viewpoint. This presentation concerns the evolving nature of representations regarding levels of abstraction and degrees of precision, the function of external representations, and specific qualities of representation in collective design. Designing is described at three levels: the organisation of the activity, its strategies, and its design-representation construction activities (different ways to generate, transform, and evaluate representations). Even if we adopt a “generic design ” stance, we claim that design can take different forms depending on the nature of the artefact, and we propose some candidates for dimensions that allow a distinction to be made between these forms of design. We discuss the potential specificity of HCI design, and the lack

software code reuse

This text adopts a cognitive viewpoint on design, focusing on individually conducted activities actually implemented in professional, industrial design projects. It presents elements for a cognitive descriptive model of design that, on the one hand, furthers our understanding of design, and on the other hand, offers elements to people who wish to use such knowledge in order to advance education and practice of professional designers. The text is especially concerned with dynamic aspects of design —that is, it focuses on the activity implemented by designers, especially the cognitive processes and/or strategies they use — rather than with static aspects. Section 1 presents the classical cognitive viewpoint on design, that is, the symbolic information-processing (SIP) approach, represented by Herbert A. Simon. Section 2 focuses on the main alternative to the SIP approach for design, i.e. the "situativity " (SIT) approach, mainly represented by Donald Schön. Section 3 is the main division of this text. It presents nuances and critiques with respect to both SIP and SIT approaches, and completes and integrates these two approaches into our own cognitively oriented dynamic approach to design.

ices Reuse strategies (horizontal/vertical) Changes to practices used in system design Underpinning technology needs /* reason to precede existing section 3 */ Component development practices Testing and Evaluation Recommended practices Documentation 3. Discussion of design practices for CBSE At present no widely recognised software design practices incorporate the concept of reusing pre-existing components. The 'historical' approach to software development has (on paper at least) encouraged the adoption of design methods such as SSA/SD, JSD, OOSD etc., which assume: A 'clean slate' solution, in which each design solution is custom-built to fit the needs of the particular problem. The reuse of design 'process' experience is mainly conveyed through the practices of the design method. Indeed, some of the current interest in the concept of 'design patterns' [Gamma et al., 1995] may well reflect the difficulty that these existing practices have in incorporating guidance

RESUME. L'utilisation concurrente de différentes méthodes de recueil de données est proposée, chacune étant appropriée pour recueillir un type particulier de données. L'exposé de son application au recueil de connaissances en programmation sert d'illustration. Dans une première étape d'étude d'un domaine nouveau, des entretiens permettent d'obtenir des informations générales sur l'organisation de l'activité et des thèmes d'étude à approfondir à l'aide d'autres méthodes. L'analyse du produit de l'activité permet une étude détaillée des connaissances que possède l'opérateur, mais ne fournit que des hypothèses sur la façon dont il les utilise. L'observation en temps réel en situation de travail donne accès à cette utilisation des connaissances, mais-pour des raisons de coût, ne permettant pas l'étude de beaucoup d'opérateurs- nécessite, en général, une validation indépendante des résultats. L'observation en situation contrôlée permet d'étudier beaucoup de sujets, mais sur un nombre de facteurs restreint et, en général, dans un contexte plutôt limité. L'utilisation concurrente de ces méthodes permet alors de neutraliser les inconvénients de chacune prise individuellement, tout en tirant profit des avantages de chacune. MOTS-CLES. Méthodes de recueil de données, Expertise, Activité de programmation, Entretiens, Observation de l'activité, Observation en temps réel, Représentation des connaissances.

Abstract not found

Problem solving based on "reuse " of problem-solving elements, i.e. particular solutions to particular problems, rather than on the use of general problem-solving knowledge ("problem-solving reuse", or "design reuse " in this text), is considered to play an important role in design 1,2,3 (cf. Case-Based Reasoning 4). This affirmation which is often made, is generally based on introspection by authors who are design methodologists, A.I. researchers or designers themselves. The conclusion which may be drawn from their papers is that designers may have good reasons to proceed to design reuse. A question which may be asked however is: do designers indeed proceed to reuse, and, if they do, why, and how do they proceed? Only a few empirical studies have been conducted on this question, nearly exclusively in the domain of software design 5,6. The study presented in this paper intended to examine the question of design reuse in order to identify requirements for reuse support tools through a characterisation of the actual reuse activities. The analysis conducted and the first results obtained led to a change of perspective: we discovered that, in addition to the reuse of problem-solving elements, the use-or reuse- of other types of "episodic " data, particularly knowledge, also played an important role in the design problem solving examined. This paper will present the questions and hypotheses formulated at the start of the study and the elements which led to modify the original perspective (first section), the results on actual probleminria-00633727,

The paper presents an investigation of the differences between two experts in the same domain. The observed differences concern comparisons between domain objects, rule justifications (technical vs.pragmatic justifications, naïve physics reasoning), and categorical knowledge (logic, level, and extension of the categorization). Differences are attributed to the prior experience of the two experts (workshop vs.laboratory). Implications for knowledge elicitation and for the design of assistance tools are presented.