RPV: Triples Made Plain describes Tim Bray's new proposed serialization of RDF assertions.
The current RDF serialization format confuses me. RPV seems much clearer. Maybe, hopefully, it'll lower the bar for RDF-enabled applications?
There's been musings of an RDF semantic module for drools and possibly hooking drools into some Linda/TuplesSpaces implementations, such as JavaSpaces or JXTASpaces. Piping space objects and properties to drools via RDF, or ideally, RPV, would definitely allow some coolness.
Hi Bob,
Yep, there have been some interesting discussions following from Tim Bray's comments about using RDF for RDDL, which took him to RPV. As it happens RDF XML syntax as it stands can be presented in a form that isn't much worse than RPV, though most of the examples you see are pretty ugly. There are other non-XML syntaxes, btw.
My own take on this is that all too often people (especially XML people) approach RDF from the syntax, which can be really complex, rather than from the model which essentially is quite simple. Whether you use RDF/XML, RPV, n3 or NTriples syntax doesn't really matter - it's effectively just a view.
The discussion on xml-dev led me to put together "RDF in 500 Words" in which I've avoided syntax altogether (no pictures either - but it was only intended for the list, I'll tart it up later)
http://www.citnames.com/2002/11/rdf500.htm
Regarding RDF & drools - I'm hoping to get back to looking at this once my latest coding cycle is done for Christmas. It looks like the best way of doing it will be through the OWL (Web Ontology) language, which is the successor to DAML+OIL and has the inference bits that drools could plug into.
Your ideas regarding object & spaces certainly do sound cool!
Keep up the good work,
Cheers,
Danny.