- From: Booth, David (HP Software - Boston) <dbooth@hp.com>
- Date: Tue, 15 Jul 2008 16:21:22 +0000
- To: Paul Prescod <paul@prescod.net>, Mark Baker <distobj@acm.org>
- CC: John Bradley <john.bradley@wingaa.com>, "www-tag@w3.org" <www-tag@w3.org>
> From: Paul Prescod > [ . . . ] > I feel like there are certain irreconcilable goals here: > > 1. use HTTP URIs (and protocol) for HTTP-only applications > > 2. add additional functionality beyond HTTP for XRI-aware > applications > > 3. encode the trigger for that functionality *in the URI* > and not in markup or elsewhere > > 4. keep URIs opaque The AWWW says: http://www.w3.org/TR/webarch/#pr-uri-opacity [[ Good practice: URI opacity Agents making use of URIs SHOULD NOT attempt to infer properties of the referenced resource. ]] But I think the words "opaque" and "infer" may be a bit misleading in the context of this discussion. When this TAG advice was written, the main concern was that some agents were looking at ".html" at the end of a URI and using that to "infer" (i.e., *guess*) that the result would be an HTML document. The important principle behind this advice is that there must be an unambiguous, well-defined chain of authority for interpreting the meaning of the URI and what it denotes, starting with the scheme. This is the same principle behind the TAG finding on Authoritative Metadata: http://www.w3.org/2001/tag/doc/mime-respect-20060412 This does *not* mean that an agent must not look at components of the URI in order to determine its meaning. It means that it must not *guess*. How a URI is interpreted must be explicitly licensed within this chain of authority. So when presented with a URI such as http://xri.example/foo.html the agent should *not* look at the ".html" suffix and *guess* that it will return an HTML document, nor should it look at the "xri." component and *guess* that the URI is really an HXRI-encoded XRI. However, when presented with a URI such as http://xri.net/xri/foo *If* the owner of the xri.net domain has explicitly said all URIs under the http://xri.net/* space are actually HXRI-encoded XRIs, then an agent does have license to interpret the URI that way. It would *not* be a guess. David Booth, Ph.D. HP Software +1 617 629 8881 office | dbooth@hp.com http://www.hp.com/go/software Statements made herein represent the views of the author and do not necessarily represent the official views of HP unless explicitly so stated.
Received on Tuesday, 15 July 2008 16:33:12 UTC