- From: Geoffrey M. Clemm <geoffrey.clemm@rational.com>
- Date: Mon, 8 Jan 2001 07:24:47 -0500 (EST)
- To: ietf-dav-versioning@w3.org
From: "Mark A. Hale" <mark.hale@interwoven.com> I guess that I keep going back to the point that there may be occasions where 'non-stable' URL's are used. These include: - Reusable temporary version URL's - Mutability concerns Please review the "variant" proposal. This gives you version-controlled resources with server defined (but human meaningful) names. The URL for a variant is explicitly defined not to be stable. Therefore a variant should both provide you with the "reusable temporary" and the "mutable" resource you require. Even in a stable URL, we may even have mutability concerns as being discussed in the current threads. You reference that the above suggestion is not very useful, and the point is that it is not useful unless you have a stable version URL and systems may have others. Please understand that I can appreciate the desire to have 'stable' version URL's from a long-lived client cache perspective. There are two very different behaviors: immutability with a stable name, and mutability with a re-usable name. Saying that these are two different "kinds" of a version is inherently non-interoperable, since most code written to work with one of these behaviors will not work with the other. If instead we define two different kinds of resources (versions and variants), and require that a server support one of them (versions), then we get interoperability because every client can count on the presence of immutable stable behavior on every server, and those servers that want the "mutable, re-usable name" behavior have a defined way to interoperate with each other (the variant option). Is there anything you need to do that is not provided by the current protocol with the "mutable version" option replaced by the "variant" option? Cheers, Geoff
Received on Monday, 8 January 2001 07:25:37 UTC