An empirical study of software reuse by experts in object-oriented design

This paper presents an empirical study of the software reuse activity by expert designers in the context of object-oriented design. Our study focuses on the three following aspects of reuse : (1) the interaction between some design processes, e.g. constructing a problem representation, searching for and evaluating solutions, and reuse processes, i.e. retrieving and using previous solutions, (2) the mental processes involved in reuse, e.g. example-based retrieval or bottom-up versus top-down expanding of the solution, and (3) the mental representations constructed throughout the reuse activity, e.g. dynamic versus static representations. Some implications of these results for the specification of software reuse support environments are discussed. 1 FRAMEWORK AND GOALS Software Reuse is currently one of the most active and creative research areas in Computer Science. This is mainly because software quality and productivity are assumed to be greatly increased by maximising the (re)use of (part of) prior design products instead of repeatedly designing from scratch (Krueger, 1989). Currently, the predicted level of reuse has not yet been reached, due to technical, organisational and ergonomic factors (Tracz,1987). This paper presents an empirical study1 of the software reuse activity. The originality of this study is twofold. First, it is one of the few experimental investigations into the software reuse activity by expert designers. Second, it examines reuse in the context of object-oriented programming which is assumed by its advocates to facilitate reuse. Few studies (Détienne, 1991; Lange & Moher, 1989) have analysed reuse in that context. 1 This research is partially sponsored by the European Esprit III SCALE 6334 project (System Composition and Large Grain Component Reuse Support) Our study focuses on the three following aspects of reuse : the interaction between design and reuse processes, the mental processes involved in reuse, and the mental representations constructed throughout the reuse activity. First, we analyse how some design processes, e.g. constructing a problem representation, searching for and evaluating the solution(s), and reuse processes, i.e. retrieving and using previous solution(s), may interact. For example, recalling solutions may lead to a revision of the currently-developed solution and retrieving a past solution may produce the addition of constraints to the representation of the current design problem. Second, we analyse some of the mental processes involved in the reuse activity. We investigate the retrieval processes involved in reuse. One issue is whether retrieval may be based on contextual and episodic cues, as shown in studies on the elicitation of category members (Walker & Kintsch, 1985) , or on abstract features, mainly used by the software community as a basis for information retrieval. We INTERACT'95, Lillehammer Norway, June 27-29 1995 2 also investigate the direction of the solution expansion on which the reuse activity is based, i.e. bottom-up versus top-down. Third, we are interested in the characteristics of the mental representations constructed during reuse. Psychological studies (Adelson & Soloway, 1988) pointed out that experts involved in non-routine design often simulated their more or less complete solution. The distinction is made between “dynamic” mental representations, i.e. involving simulating and enacting of a mental model, and “static” mental representations, i.e. involving static statements such as states and properties. We assume that designers may favour one such representation rather than another depending on the task involved (i.e. reuse versus design).