On Thu, Feb 19, 2009 at 2:19 PM, Anselm R Garbe <garb...@gmail.com> wrote: > The resource access should be done using a stateful protocol rather than HTTP.
I'd be very cautious about making things have an inherently stateful protocol rather than making then (a) finegrained (possibly designed to have multiple "queries" in a single network packet) and (b) able to query for local "data"/"resources"/whatever which might be cached. This is partly because in my experience with flaky 3G networks and dodgy WiFi access (which a _totally general_ HTTP replacement must handle), dropped connections happen an awful lot. I think with (a) and (b) you get most of the advantages of statefullness without the big problems if things must be stateful. (Does anyone actually know how something like 9P works on a really, really unreliable network connection?) (It also means you can try new approaches for data where the entire stream needn't be "completely reliably received", eg, network audio or video packets.) -- cheers, dave tweed__________________________ computer vision reasearcher: david.tw...@gmail.com "while having code so boring anyone can maintain it, use Python." -- attempted insult seen on slashdot