WSCOLAB: Structured collaborative tagging for Web service matchmaking

A key driver for Service Oriented Architecture implementations is the hope to save development time and costs through a higher degree of reuse of readily deployed Web services. To achieve this goal, it is necessary to make services discoverable. Current approaches use authoritatively defined taxonomies to categorize services, which simply does not work for the flood of services being published on the Web in uncoordinated manner. Collaborative tagging claims to address this problem, because it is the process by which many users add metadata in the form of keywords to shared content (Golder and Huberman, 2006). However, it suffers from the lack of structure to describe Web service interfaces. We introduce structured collaborative tagging classification schema that can provide Web service descriptions for both categorization and interface matchmaking approaches. We report an experiment where 27 taggers have used our classification schema to annotate 50 Web services from Jena Geography Dataset. We also propose a single-request Web service matchmaking technique based on the proposed classification schema. The performance of the proposed technique has been confirmed by the first position awarded in the Cross-Evaluation track of the Semantic Service Selection 2009 contest against other matchmaking approaches. It shows that our classification schema is simple and powerful enough to be used by both tagging and querying end-users.