- From: Sandro Hawke <sandro@w3.org>
- Date: Sun, 19 Mar 2006 20:48:35 -0500
- To: John Cowan <cowan@ccil.org>
- Cc: Dick Hardt <dick@sxip.com>, Larry Masinter <LMM@acm.org>, 'Dan Connolly' <connolly@w3.org>, 'John Merrells' <merrells@sxip.com>, 'Lisa Dusseault' <lisa@osafoundation.org>, uri@w3.org, 'Thomas Roessler' <tlr@w3.org>, dix@ietf.org
> Dick Hardt scripsit: > > > To elaborate, anything starting with dix:/ would be reserved for > > identifiers in the DIX standard. eg. dix:/core#1 > > Anyone with a domain name can define their own. eg. dix://acm.org/foobar > > It looks like the tag: URI scheme would be appropriate: it's used for > identifying > rather than locating, and anyone with a domain name or email address > can assign one. Indeed, if it's important to have the URIs not be dereferenceable, tags URIs (RFC 4151, which I co-wrote) are a fine choice. On the other hand, it's often very useful to provide information about the identified thing, via URI dereference, so I strongly recommend using an HTTP one, even though it can feel like like a strange approach. -- sandro
Received on Monday, 20 March 2006 01:48:47 UTC