[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: PaceErrVerb


The consensus at present on the ietf-http-wg@xxxxxx mailing list seems to be that
a header pointing to the error service is a good idea.

Henry

On 23 Jun 2004, at 21:01, Walter Underwood wrote:


--On Wednesday, June 23, 2004 02:44:16 PM -0400 Robert Sayre <mint@xxxxxxxxxxxxxxx> wrote:
+1. Atom should not need to extend HTTP. -Tim

Do you consider extension headers unacceptable as well?

I do. We should use HTTP, not re-engineer it.

First: GET, POST, and URL parameters are widely-understood techniques.
They have been adequate for thousands of applications. CGI is good
enough for Atom. Using new methods or new headers makes it harder
for implementors. If we can't make Atom work with CGI, it is our fault,
not HTTP's.

Second: If Atom requires an upgrade to the entire WWW infrastructure
(firewalls, caches, servers), it will never be adopted. A new HTTP method
will require that sort of upgrade. Are you ready to upgrade all of AOL?

wunder
--
Walter Underwood
Principal Architect
Verity Ultraseek