Advanced Web API search patterns adding collective knowledge to public repository facets

Abstract

Nowadays web designers are able to perform fast development of web applications by aggregating Web APIs available in huge public repositories. This approach is referred to as web mashup. Web APIs may be searched and aggregated in different scenarios, namely the development of a new mashup, or the completion of an existing mashup, or the substitution of one or more Web APIs within it. Collective knowledge of other designers who used the same or similar Web APIs in their own mashups is of paramount relevance to support Web API selection. In this paper, we provide a framework of techniques based on such a collective knowledge to support Web API selection according to the different search patterns, which correspond to distinct mashup development scenarios. We also present a preliminary evaluation of the framework.