- From: Mark Baker <distobj@acm.org>
- Date: Fri, 30 Jan 2004 15:59:14 -0500
- To: Mario Jeckle <mario@jeckle.de>
- Cc: www-ws-arch@w3.org
On Fri, Jan 30, 2004 at 09:15:02PM +0100, Mario Jeckle wrote: > +1 Great, thanks. > Although, REST may be sufficient for some Web Services it is not > sufficient for all (consider intermediaries, header processing, etc.). Really? REST is *extremely* constraining regarding how header processing is performed. I can't imagine a more robust header processing model than the one which REST suggests (i.e. a self-descriptive one). Don't confuse it with HTTP's model, which is not RESTful in a handful of important ways. In fact, SOAP's header processing model is far more RESTful than HTTP's. I believe intermediary support follows naturally from a solid processing model (including headers) too, though I would say that further constraints are necessary on top of REST to enable more flexible forms of intermediary composition & routing. See ARREST for some of those; http://www.ics.uci.edu/~rohit/DecentralizingREST.pdf Mark. -- Mark Baker. Ottawa, Ontario, CANADA. http://www.markbaker.ca
Received on Friday, 30 January 2004 16:00:34 UTC