Danny Ayers posted to the cmsblogapidev mailing list today in response to a question posted to the list: “Does anyone have strong opinions about databases here? if so, do they extend beyond uses for mysql?” His reply:
I don’t know about it being a strong opinion, but I believe in the context of cms and blog APIs, the relational model that is behind mysql and most other databases used for these purposes is a poor match for the requirements.I think the same problem exists with XML and object databases.
The data dealt with by cms and blog tools can be better described in terms of web resources, so the obvious choice for the modelling would be the Resource Description Framework. Ok, so RDF stores are currently few and far between, and many are actually backed by RDBMSs. But in any cms or blog system there will be the need to model the data, and IMHO it is better to reuse existing standards wherever possible (i.e. let the RDF store deal with
The use of RDF will also simplify extension of such a system, thanks mainly to use of RDF vocabularies (RDFS). As a simple example, adding FOAF (friend of a friend) capability to a system designed to handle RSS data based on a relational DB would involve building a whole new set of model constructs before anything else could be done. Adding it to an RDF-backed system would simply be a matter of setting up the appropriate user interface.
I can’t think of a better point at which to announce a new project, JemBlog,
which aims to create server-side web blog system along the lines of the core of Movable Type etc. The difference being that the model used behind the scenes is RDF, so true semantic blogging is possible. The system will be written in Java/JSP, using the Jena RDF toolkit.
This is a new project, so there’s nothing usable available yet. I’m hoping to have the first working draft together in the next couple of weeks. This is an open source project (probably LGPL) so input is welcome from everyone.
There’s a sourceforge mailing list and download archive.