Raw

This is kinda rich, Danny, since I was *the* protocol maximalist on DAWG; that is, I wrote a draft of the protocol spec that did exactly (more or less) what you talk about here: used GET, PUT, POST, DELETE, and OPTIONS; had update; had graph creation and maintenance methods (drop & create graph -- this was *before* the QL got named graphs, btw), etc.



But, guess what? DAWG is (or was then) full of protocol minimalists, I fought a good fight, and lost *across the board*. So, we get a very minimalist SPARQL Protocol because that's what the WG wanted.



I'm not happy about that, but the process is the process.



As for update, it's out of scope, per the charter. I didn't write the charter and think it's a pretty bad example of the art, but, again, there was no consensus on how or whether to alter the charter in the WG.



Two final points: "Whatever happens, I think opaque tunnelling a la XML-RPC/SOAP should be avoided like the plague" is a curious comment; surely yr not directing that at SPARQL Protocol? It does no such thing. It's a pure HTTP, REST-friendly thing. It uses WSDL 2 to describe that HTTP protocol, but that's rather a neat trick in my book and has NOTHING whatever to do with XML-RPC or SOAP (that the protocol also has a SOAP binding doesn't make it less REST-friendly, it just gives people other options).



This is all interesting because both "SOAP" and "web services" are on the list of things that were going to "break the Web"... Why don't we modulate our view of the utility of Mark Baker's fundamentalism when his predictions that SOAP and web services would break the Web turn out to be *amazingly wrong*?



Finally, Mark is just *wrong* on the ESW site when he claims that "the current notion of SPARQL protocol bindings is broken". If you think that Tim is god and Roy Fielding is his prophet, then there probably is some problem here -- though why Tim, with his powerful ability to veto stuff, let the Web Services crowd risk 'breaking the Web' is beyond me...



But Mark's feedback was heard, responded to, and considered at length on DAWG. I personally spent many hours doing that. We just weren't convinced. I'll put it this way: as a matter of empirical fact, the Web is hard to "break", and the dire predictions otherwise have all been *wrong*. If the Web could be broken by misusing HTTP, it already would have broken, but, more to the point, it wouldn't be worth having.@en