Raphaël Luta, mentor of the incubating Apache Graffito project, recently asked for volunteers to help him mentor the Graffito project. Encouraged by my recent ASF membership and the fact that I had been keeping an eye on the project for quite a while, I decided to volunteer.
Before formal appointment as a mentor I’ve introduced myself on the Graffito mailing list and started getting more familiar with the project. I’m especially interested in the Object/Content Mapping (OCM) framework JCR-mapping that is included as a Graffito subproject, but also on the other parts of the project. Having worked on various content management systems for over ten years since our first pre-Midgard site experiments (including a custom SGML vocabulary mapped to early HTML with DSSSL) with the Greywolves, I’m still eager to learn new approaches like Graffito.
Graffito is both a framework for building content applications and a set of existing portlet components. Although the project is related to Apache Jetspeed-2, the portlets should work on any compliant portal engine. The framework components can also be used outside the portlet model. The project aims also for independence from the underlying content storage, using a generic “Graffito Core Model” and itsderivates as the abstract content model and “Content Server” abstraction for the content storage layer.
The current default content server implementation is based on Apache OJB and runs on top of relational databases. The JCR-mapping subproject is planned to be used for a similar task on top of JCR content repositories, especially Apache Jackrabbit. Just like OJB is not a subproject of Graffito, we’ve had initial discussion about possibly moving the JCR-mapping project outside Graffito. Making it a Jackrabbit subproject is an interesting alternative, especially if we want to target for an eventual JCR federation within the ASF, but for now I think it’s best to get the communities together and see what patterns emerge before making final decisions on what to do with thesubproject.