I added some examples to the query page myself, to help people get an idea, but I am happy that you did too. I chose to build upon Redland/Rasqal because it fits rather well into the existing architecture, and allowed me to get this out rather fast, as it is indeed something that will allow people to use these data in novel ways and so bring more interested users to Opera and our community.

It is just a start, allowing experimentation and allows us to gain experience in a field where there is little experience to build on. Clearly, as usage soars, much more work will be needed.

I do not anymore allow the retrieval of the full model, as I rewrote the resource limiting code to emphasise memory use rather than a timeout, and now a too big or complex query will result in an error. Unfortunately, it will currently result in a proxy error rather than a internal server error, I will have to fix that later.

Also, I'll try stay active in the Semantic Web community, so it shouldn't be necessary to reverse engineer the model, just ask! :-) I intend to follow the work on service descriptions closely, but time doesn't allow to elaborate on this right now. Briefly, our user data is modelled mostly using FOAF, our galleries using FOAF Image and my own Gallery schema, while blog posts and forum contributions are mostly Dublin Core.

So, we're just getting started, but I hope it will also get the Semantic Web and the Web 2.0 communities started!