A Semantic Framework for Collaborative Enterprise Knowledge Mashup

In this paper, we propose a semantic framework to support enterprise mashup within or across collaborative partners. The aim is to enable effective searching and finding of mashup components and their composition, by making possible proactive suggestion of mashup components and progressive mashup composition. The framework is constituted by a model of component semantic descriptor, apt to abstract from the heterogeneity of underlying APIs, and by techniques for building a mashup ontology where semantic descriptors are semantically organized according to similarity and coupling links. The semantic framework can be exploited to support an exploratory perspective, where the user has not exactly in mind what is the mashup application to build, but new components are suggested on the basis of their similarity or coupling with respect to already selected ones.