The Role of Similarity in the Reuse of Object-Oriented Analysis Models

• Authors: Gretchen Irwin
• Venue: Journal of Management Information Systems
• Year: 2002

Software reusethe application of existing software artifacts inthe development of a new systemhas been claimed to dramaticallyimprove systems development productivity and quality. These claimshave been particularly pronounced with respect to the reuse ofobject-oriented (OO) software artifacts. However, the empiricalevaluations of these claims are relatively sparse and ofteninconsistent. This paper begins to address the gap in theliterature. A verbal-protocol study was conducted in which analystscreated a model for a problem (the target) and were given anexample problem and solution (the source) to reuse. Theresults show little support for reuse in OO analysis. First, reusehad no effect on the quality of the OO analysis models. Subjectsgiven a highly reusable example produced solutions that were nobetter than those of subjects in the control group. Second, thedegree of similarity between the source and target problems didhave an effect on the reuse process, although it did notimpact the reuse outcome. Subjects given the example withthe most similarity to the target problem quickly recognized thereuse potential, attempted a fair amount of reuse, but made severalerrors stemming from lazy copying. Subjects given an example with alesser (but still significant) degree of similarity were oftenunable to recognize the reuse potential, and thus engaged in lessreuse activity. Thus, the characteristics of the source-targetcomparison that facilitate noticing the reuse potential ofthe source do not necessarily help in applying the sourcesolution to the target problem. These results suggest that theclaims associated with reuse should be treated with a healthy doseof skepticism.