2009/7/13 Erik Wilde <[email protected]> > Mark Nottingham wrote: > >> Personally, I'm getting sick of the "can't set HTTP headers" scenario >> driving abuse of HTTP. >> > > just for the record: i was not proposing to re-create HTTP headers as atom > extension. i was suggesting that maybe some information about the update > properties of a feed might be useful, and my guess is that these might go > well beyond the properties expressible in HTTP. but i do agree that if all > we want to do is what HTTP can do, then we should definitely stick with HTTP > headers. > > cheers, > > dret.
I was aware that there might be a different application protocol used for transport; e.g. Jabber sending entries about, but I didn't want to be too prescriptive when asking the OP for clarification. It might be worthwhile discussing what information gets lost when HTTP is not involved. The initial statement talked about clients polling a publisher, which lead me to assume HTTP was being used. I'm with Mark in that people using the web should *use* the web, and think about Last-Modified, If-Modified-Since and Expires, or ETag, If-None-Match and Cache-Control. That's not finger-pointing, but merely an indication of the types of services that I would prefer to consume. Cheers, James
