Developers working on a large open-source software project like Eclipse, face the problem of accessing knowledge accumulated during development work on the project when they are separated by the geographic distance or time. We present a system that ensures that this knowledge is accessible to all team members, through the formation of a group memory from the artifacts created as part of a software development. This group memory is formed automatically by inferring links between archived artifacts, such as the source code, email messages between team members, and problem reports. A part of the system is a recommender that suggests pertinent parts of the group memory to a developer engaged in software evolution tasks.
The demonstration will present a working prototype of our system that runs in Eclipse supporting Eclipse developoment. We will run through several typical usage scenarios with Hipikat, addressing increasingly sophisticated Eclipse developer audience.
These will range from searching the artifact database for explanation on programming a certain feature in Eclipse API to a usage scenario with an Eclipse developer working on an open Bugzilla item. The recommendations - from tutorial articles on the Web site to similar items already in Bugzilla - will then be further followed to related artifacts like newsgroup postings, Eclipse code fragments implementing and/or using a feature, and files changed to implement particular bug fixes, focussing the user on useful classes and relevant APIs.