Profiting from case-based reasoning in framework documentation

The main weakness of frameworks is their steep learning curve. So, improving their documentation is a critical issue. We present a case-based reasoning approach that complements existing documentation methods and faces their main drawbacks. The main contributions of our work are: the development of an approximate retrieval mechanism to locate the most similar framework usage experience (case) for a query posed by the user. That way we cope with the typical sequential searches in cookbooks, very inefficient in a real context. Secondly, the development of an interactive adaptation mechanism that guides the user in obtaining new usage experiences. Thus, we face the fact that it results unrealistically in trying to represent every use of the framework. Thirdly, the existence of a learning mechanism that lets the user enrich the case base with new experiences, reducing the initial effort of the usage experience acquisition. Our approach is illustrated by OoFRA, a prototype tool that provides intelligent documentation for a framework in the graphical user interface domain.