- From: Henrik Nordstrom <henrik@henriknordstrom.net>
- Date: Tue, 18 Mar 2008 19:51:04 +0100
- To: Mark Nottingham <mnot@mnot.net>
- Cc: Lisa Dusseault <lisa@osafoundation.org>, Robert Siemer <Robert.Siemer-httpwg@backsla.sh>, Werner Baumann <werner.baumann@onlinehome.de>, ietf-http-wg@w3.org
On Tue, 2008-03-18 at 09:44 +1100, Mark Nottingham wrote: > This crossed my mind as well... Weak ETags are in use today, but can > we find a situation where they're actually improving things, and > getting interoperability? There is no real interoperability problem with weak etags today, other than some servers not implementing weak matches in their default filesystem based conditional processing. Clients behave well, and servers using (not only sending) weak etags can make good use of them. The only real problem is this discussion, of which maybe 80% is fud from not understanding how servers use weak etags today and people drawing far going conclusions from just a small part of the picture, 10% misunderstandings of what weak etags may be used for, 5% total confusion and 5% actual discussion moving forward. Personally I wish we could extend weak conditionals to be allowed in If-Match processing for most methods which would make the protocol more complete for the situations where one need (or wants) to use weak etags, but thats outside our scope. For whats within our scope the existing definition of weak etags and weak conditionals is very sane, if one ignores the confusion about what is actually meant by "semantic equivalent" which is a product of the language used in the descriptive text outside of the actual definitions combined with those 10% who wants to use etags for more than they are (part of that also visible in the meaning of ETag in response to PUT discussion). Getting rid of the "semantic equivalence" mistakes, sticking to "no significant semantic difference in point of view of the server" whould help a lot in keeping people on track on what a weak etag is and meant to be used for. Together with a reminder that the only 2616 operation allowing a weak condition is GET If-None-Match, which for most uses of weak etag is quite sufficient, but falls short for extensions like WebDAV or CalDAV who have additional needs (CalDAV in particular as it's application scope makes strong ETags quite unsuited, with a lot of not fully deterministic server-side processing going on affecting the octet equivalence) Regards Henrik
Received on Tuesday, 18 March 2008 18:53:11 UTC