Retrieving Reusable Software Components Using Enhanced Representation of Domain Knowledge

This paper describes an ontology-based approach for identifying and retrieving relevant software components in large reuse libraries. Since it is usually difficult to precisely identify exact matches without considering domain knowledge, we exploit the use of domainspecific ontologies to enrich a knowledge base initially populated with multi-faceted ontological descriptions of API components. In addition to pure semantic-based search, this enriched knowledge base supports signature-based search, keyword search, and blended search. However, our experiments show evidence that only semantic search that is backed by reasoning services enables intelligent matchmaking and yields improved precision. Based on a usability case study, we further argue that semantic search is indeed usable and practical.