Deriving folksonomies for improving web API search

Abstract

Web APIs, that is, software components made available by third parties through web interfaces, can be aggregated to develop web applications, also known as mashups. Also in this application domain, tagging performed by other mashup designers, who used available Web APIs and mashups composed of them, might be exploited as knowledge that progressively emerges from the community of designers. Web API tagging has some peculiar aspects that will be analyzed in this paper. On the other hand, folksonomies are Web 2.0 tools for conceptualizing knowledge emerging from the bottom. In this paper, we discuss the adoption of folksonomy concepts in modeling Web API use for mashup development. We motivate the adoption of folksonomies in this context and we present the differences with other models that represent very close information. Our folksonomy model is meant to be fully compliant with existing and commonly used public Web API repositories. It is not intended to substitute them, but to complement their contents in order to enable advanced Web API search facilities in such a collaborative environment.