- From: Lisa Dusseault <lisa@osafoundation.org>
- Date: Sun, 17 Oct 2004 15:52:11 -0700
- To: Julian Reschke <julian.reschke@gmx.de>
- Cc: ietf-http-wg@w3.org
Jeffrey Mogul suggested this change from the -05 model; I think it works at least as well as the model where the request body is modelled as an instance. It's entirely consistent with RFC3229's model. And instead of using the IANA MIME type registry, it uses the IANA instance-manipulation registry. Do you have any technical objections to using the IM header and the instance-manipulation values? On Oct 17, 2004, at 2:47 AM, Julian Reschke wrote: > > That seems to break HTTP semantics, because the Content-Type header is > supposed to identify the type of the request body, not the one of the > entity it's applied to. Furthermore, the information seems to be > completely useless unless you specifify what a server is to do in case > of a mismatch. > The model of HTTP used by RFC3229 is one in which the Content-Type header applies to the entity's body type. Then when an instance manipulation is applied the request body might end up in gzip format or in a patch format, but the Content-Type doesn't change. The proposal does specify what to do in case of a mismatch -- the server is requested to change the MIME type of the resource to the new one provided by the client. BTW, the change to call the patch documents "delta encodings" is also to be more consistent with RFC3229 now that I understand it better. I really don't know what to do about the IPR issue around VCDIFF since I am not a lawyer; since the IESG already made RFC3284 into a Standards Track document, I assume this will be acceptable. Thanks, Lisa
Received on Sunday, 17 October 2004 22:52:30 UTC