- From: Larry Masinter <masinter@parc.xerox.com>
- Date: Tue, 12 Nov 1996 21:04:23 PST
- To: yarong@microsoft.com
- CC: kt@nttlabs.com, w3c-dist-auth@w3.org
# We need to be able to lock, copy, move, redirect, delete, destroy, etc. # attributes just as we can do the same with any other resource. Your # syntax does not allow this. The metadata for a resource is just another resource. The 'link' attribute notes the relationship between the URL for the resource and the URL for the metadata of the same resource. I prefer a dynamically retrieved 'link' attribute rather than one where the derivation is made syntactically by the client. I think it is good to let URL syntax be opaque whenever possible. However, I don't think that metadata should be treated as a set of independent resources, except perhaps as a compound object. That is, normally you can think of 'the metadata for the message' as a single object. It might be a compound object and let you manipulate 'title' independently of 'author', but generally they go together in a set, and the cases where they allow independent manipulation are no different than those where a system will let you update 'page 12' of a document independently of 'page 13'. Larry
Received on Wednesday, 13 November 1996 00:04:28 UTC