On Thu, 04 Nov 2004 10:28:49 -0500, Robert Sayre <[EMAIL PROTECTED]> wrote:
Personally I'm still unsure how ERROR would be any less responsible than using POST, but even if I were convinced, then there are still other issues. > > inventing new HTTP verbs is really questionable. > > That's the subtext of all the arguments for POST, I think. > Unfortunately, it's not really backed by anything other than religion. Religion isn't idempotent. The more people follow the religion, the less the chance of your (heretical) suggestions being accepted. The orthodoxy at the moment is that adding HTTP verbs is a bad idea. Whether this underlying belief is true or not, it renders it virtually impossible to add HTTP verbs across the board (c.f. URIQA). A more immediate problem is the need to make modifications at the HTTP layer. That has practical implications, in particular as regards who will be *able* to implement. > It's also humorous to me that most of the people who feel that way have > no problem with inventing new verbs by extending HTTP with custom headers. Hmm, arguably the custom headers act more like adverbs or adjectives, but there are admittedly similar implementation issues. Cheers, Danny. -- http://dannyayers.com
