Pre-filtering repository models

Software reuse is the use of existing software artifacts to build new system rather than building from scratch. It has potential in improving the quality, the productivity, and reduces the overall software development cost. Reusable software components are usually kept in a components library or repository, from where they can be retrieved during reuse. In this paper we present a fast way of identifying a subset of repository models that are potentially similar to a user query prior to retrieval stage. The shortlisted repository models are then compared with the query models in a subsequent computationally demanding retrieval stage to ascertain their actual degree of similarity with the query models. Our method of selecting the subset of repository models is called Pre-filtering. This technique lead to significant reduction in retrieval time, especially when the repository contained large number of candidates solutions to be analyzed during retrieval.