Capitalizing the Designers' Experience for Improving Web API Selection

摘要

Solutions for supporting Web API selection may depend, behind the compliance of available Web APIs with respect to the search request (according to user-specified tags, Web API technical features, such as protocols or data formats, Web API categories), on votes assigned to Web APIs by web application designers who used them in the past. A web application designer, who is looking for a Web API for his/her own mashup, may learn from Web API rating performed by other designers. Votes should be properly weighted considering the experience/skill on application development gained by designers who assigned the votes. Therefore, estimating this experience is crucial for exploiting it in the best way. In this paper, we propose new techniques for the estimation of this experience by combining several factors, such as the reputation and popularity of the applications developed by the designers in the past. We validated our proposal with preliminary experiments, based on the contents of a well-known Web API public repository.