Hi Kendall, thanks for the responses.
Ok, I guess I'd better find a better phrase than "breaking the Web". Maybe just "abuse" ;-)
In regards to SPARQL as of now - it currently seems to respect WebArch pretty well as far as I can make out, but that's not difficult when all you have to do is GET. With update operations that could easily change, there's always a temptation to cram everything into POST. For example, in the notes of the Wiki now, there's the suggestion of atomic ADD+REMOVE messages. Seems like that can't help but overload a verb. The whole thing doesn't need to be in one payload to ensure reliability of the operation.
I remember your protocol proposal - as far as I recall, it did look good (must Google :-)@en
Ok, I guess I'd better find a better phrase than "breaking the Web". Maybe just "abuse" ;-)
In regards to SPARQL as of now - it currently seems to respect WebArch pretty well as far as I can make out, but that's not difficult when all you have to do is GET. With update operations that could easily change, there's always a temptation to cram everything into POST. For example, in the notes of the Wiki now, there's the suggestion of atomic ADD+REMOVE messages. Seems like that can't help but overload a verb. The whole thing doesn't need to be in one payload to ensure reliability of the operation.
I remember your protocol proposal - as far as I recall, it did look good (must Google :-)@en