As of April, 11th, I've been mainly assessing code reuse scenarios, both for the ontology modeling and the repository design tasks; I've been working on the domain model, analyzing Maven repository's tag cloud, and getting a grasp on the OSGi specification as well.
A good axiomatization of the software component domain is given here @ cos.ontoware.com, and I think augmenting such ontologies would be nice for the project. They provide a quite granular description of a software component and library, which will support future development of the project, pushing formal description of software components down to interface- and method-level. I believe some sort of integration with TagSEA (or other efforts in the code annotation area) is possible and promising for the future.
I also discussed with my mentor some issues about OSGi, Maven and other provisioning systems seamless integration in the framework-to-be, since they share most of metadata and semantics about components: it would be nice to have a core set of shared essential metadata.
I am currenty thinking - and discussing with my mentor - about the chance of extending an existing OSGi Bundle Repository implementation, to add the needed semantic services, avoiding to design the repository from scratch (thanks lemmy). The repository could be deployed as a bundle itself. However, this requires me some time to be assessed more thoroughly.
Repository coding will start thru next weeks, as soon as I decide existing components' and technologies' integration, analyze use cases and define a strategic architecture to rely upon.
I keep updating my project's wiki quite frequently; I just added identified use cases, outlines and specifications will follow asap.
Feedback, criticism and suggestions are welcome ;) .
No comments:
Post a Comment