(A) social network-based framework for searching reusable software assets

Abstract

Software reuse is the concept of reducing the cost and efforts of developing new software by reusing existing software assets. Much research is being done on software reuse because of its importance and effectiveness. However, problems still exist. One of the biggest problems is the difficulty of finding useful software assets. More specifically, a developer does not know where to find, how to select, and how to reuse reusable assets. The main causes of these reuse problems are limitations on distribution repositories by distribution organization; limitations of existing search methods, such as facet, keyword, and semantic search; and insufficient description of reusable assets. To solve these problems, we propose a social network-based search method as a new search method to be used in addition to existing search methods. To enable developers to more easily locate reuse assets, three important aspects of developers’ social networks are defined in terms of reuse: similarity, familiarity, and favorability. In addition, eight important factors that make up each aspect are defined: common application domain, common role, common language, reuse relationship, interaction for reuse, amount of reusable assets owned, frequency of getting reused, and reputation. Furthermore, reuse knowledge to supplement insufficient description of reusable assets are defined. A social network-based search model is defined according to these three aspects and eight factors. We implement a prototype for the social network-based search model and evaluate this model using real data in the open source domain. This research may enable developers and their mangers to enhance their searches for reuse software assets and increase reusability itself by connecting developers with each other and sharing knowledge.