The Amos Project: an Approach to Reusing Open Source Code

Building reliable software products based on components whose properties are well established and understood is one of the key goals of component-based software development. There are some cases in which this approach has to face practical problems: for example, when code which has not been formally developed, but which is interesting, is needed, or when there is a very large number of components. We describe the basis and current work in Amos, an IST-funded project that aims at facilitating the search and selection process of source code assets in order to ease its reuse. We show how, although the description of source code packages is generated using a non-formal method, the search and assembly process has a sound basis. We argue that lessons learned from this approach will be of use to learn about the characterization of software code not formally developed, and about the search for the right component amongst thousands of other components.