Powerful RDF

one of the best resources available from the W3C are their Technical Architecture Group (TAG) findings. they talk a lot about design principles and architecture, and not so much about specific technologies. they really should be advertised more openly, so that people looking for architectural guidance can find them more easily.

the one that recently caught my attention is The Rule of Least Power, published in its last revision almost two years ago. in the abstract, this document states that the Rule of Least Power suggests choosing the least powerful language suitable for a given purpose.

what i find a bit surprising is the fact that the document talks about the semantic web and RDF as satisfying the rule of least power. to me, it always appears as if it is the exact opposite, a semantic framework designed to capture any semantics that anyone ever might want to express, and to be able to reason about them. that sounds pretty powerful to me. and RDF needs OWL and custom query languages and databases and a whole suite of new technologies to be fully deployed. which sounds again like a lot infrastructure that is required for a pretty powerful idea.

jeni tennison just blogged about RDF and observed:

It should be that mashing up RDF is easier than mashing up random JSON/XML. Just a few this is equivalent to that assertions and you're good to go. So show me the places (outside academia) where RDF-based mash-ups are being made. Show me the interfaces that are allowing people in the real world to use RDF to pull together information from diverse sources, get better overviews, draw more accurate conclusions. I want to believe. Show me the uses of RDF that will convert the sceptic in me.

when RDF came on stage with its claim to once and for all solve the problem of declarative semantics, it sounded very powerful, suspiciously like something we heard 25 years earlier, and pretty ambitious. 

in the meantime, the research community around RDF has become huge, and no doubt some pretty impressive applications have been created. the question i am asking myself is how much better and more cost-effective these applications are than applications which would have been created with a less powerful toolset, and whether that was really worth the effort.

i don't think in this case it is completely accurate to quote the semantic web as following the rule of least power only because it is declarative. it is enormously powerful, and at a pretty high cost, and it still has to prove that the results are worth the investments.