The distiller for RDFa 1.0 has been around for quite some time now, and has been widely used. With the advance of RDFa 1.1, I have made a fairly extensive re-write of the distiller, though, which is now available as a “shadow” RDFa 1.1 distiller. “Shadow”, because it is still beta and it does not (yet) replace the current, “official” distiller. Furthermore, RDFa 1.1 is still in evolution and hence the shadow version may still change.

Apart from the obvious changes in implementing the new RDFa features (profiles, @vocab, usage of @prefix instead of @xmlns, possibility to use full URI-s everywhere, etc.), the package has also been restructured and some old problems have been taken care of. For example, the current code looks at the HTTP return message for the content type, and decides whether the code is to be run through Python’s XML parser or through the HTML5 parser accordingly; the character encoding field of the return message is also taken into account to set the right encoding when parsing, etc.

There is also a validator service now. Although this looks like a separate service, it is, in fact, just a shell on top of the same distiller code. Indeed, RDFa 1.1 cleanly defines an error mechanism by returning RDF error triples as part of a separate graph; the only job of the validator service is to interpret these triples and display them on the screen in a human readable way. Hopefully this may become useful, too.

And a blatant self-promotion: I will give a presentation on RDFa next week, at the SemTech conference in San Francisco…