Semantic-based Planning of User Interface Mashups
User interface (UI) mashup platforms enable development of composite applications out of autonomous building blocks, called widgets. A widget is an independent re-usable Web application fulfilling a specific simple task. By combining the right widgets, users with little programming skills should be empowered to build complex solutions meeting their IT needs on their own.
However, mashup development is a non-trivial task. One has to find, configure and connect appropriate widgets to achieve the desired functionality. The composition process is even further impeded by the fact that widgets’ documentation is rather scarce and their APIs are highly heterogeneous.
The goal of the project is to analyze, to which extent the composition of UI mashups can be performed automatically. Based on a state-of-the-art analysis of semantic modeling techniques and planning algorithms, the project should develop a reusable framework for automatic composition of UI mashups. The application of the framework should be demonstrated in the context of an existing UI mashup platform such as Apache Rave.