- From: Patrick Stickler <patrick.stickler@nokia.com>
- Date: Tue, 29 Mar 2005 12:42:52 +0300
- To: "ext Tim Berners-Lee" <timbl@w3.org>
- Cc: David Wood <dwood@mindswap.org>, www-tag@w3.org, public-swbp-wg@w3.org
On Mar 27, 2005, at 22:20, ext Tim Berners-Lee wrote: > > Could the SWBPWG please answer also answer the following: I'm not a member of SWBPWG, but I'll hazard a response (even though I don't see how this moves us closer to closure of this issue)... > > 1. Who was the creator <http://www.w3.org/2005/moby/dick> ? If you mean, who was the creator of the resource identified by the URI "http://www.w3.org/2005/moby/dick", and if we trust the information accessible via that URI, then we might conclude, based on the representation returned, that the URI identifies a "book" and that a person by the name of "Herman Melville" was the creator of the "book" identified by that URI. But that is just an educated guess, and possibly incorrect. After all, the content of the representation returned could be inaccurate, or even contain one or more deliberate falsehoods. If you mean, who was the creator of the URI itself, then that information is not reliably available in any information accessible via that URI, though one might presume that the creator was the owner of the domain www.w3.org, or some authorized agent of the domain owner. We might even guess that the creator of the URI was you, and probably be correct. But again, that might also be quite wrong. It's a guess at best. > > 2. What is the year of creation of <http://www.w3.org/2005/moby/dick> ? If you mean the "book", then no such information is accessible via that URI. One might utilize the information accessible via that URI to research an answer via some other sources of information, presuming that the resource in question, the "book", is actually unambiguously identified. If you mean the URI itself, then, again, technically, no such information is available -- though one might opt to exercise bad practice and conclude from the lexical nature of the URI itself, and perhaps some proprietary knowledge of W3C naming conventions, that the URI was created in 2005. But again, that would be nothing more than a guess. If you mean the representation accessible via that URI, then the HTTP headers at least suggest that the representation was created prior to the Last-Modified: date of "Sun, 27 Mar 2005 18:54:36 GMT", which is probably reliable information -- but the actual creation year of the representation remains unclear. > 3. Who was the creator <http://www.w3.org/2005/moby/xyz> ? No idea. The representation returned appears to be a poem, so perhaps the URI identifies a poem. But then, it might identify a web page containing a poem, or a particular translation of a poem, or not a poem at all, but perhaps it is a representation of the state of a horde of drunken elves firmly convinced of the acceptability of using URIs without hashes to identify RDF properties (who knows... ;-) > > 4. What is the year of creation of <http://www.w3.org/2005/moby/xyz> ? No idea. -- Now, having gone through that "exercise"... None of the questions above are questions that can be reliably answered at the web layer -- because the web layer simply does not care what those URIs actually identify and provides no specific machinery or functionality whatsoever to communicate what those URIs actually identify (true, the representations accessible via those URIs may provide answers to those questions, but typically will not). The web layer certainly *does* care that those URIs identify *something* and that there are accessible representations of those *somethings*, and humans will care about the utility of the representations accessible and consistency of representations (but most humans won't care either what the URIs acutually identify). Furthermore, the very form of the questions introduces ambiguity as to what you refer to -- i.e. the resource identified (denoted) by the URI, or the information accessible via that URI, or the web experience associated with that URI, or any indirect subject or concept related to any of the above, etc. etc. It is simply not possible to know what resources those URIs actually identify by dereferencing them -- per traditional HTTP methods -- and any attempts to answer your questions are guesses (perhaps intelligent guesses, perhaps correct guesses, but guesses nonetheless). And none of those guesses are made any more reliable or accurate by whether (a) the URIs contain or don't contain hashes or (b) the URIs are or are not constrained to identify "information resources". -- While some members of the SWBP WG might choose to attempt to dereference the above URIs and make intelligent guesses about the most likely answers to the questions you pose, such an activity would have no relevance to the position that URIs without hashes MAY identify RDF properties. Regards, Patrick
Received on Tuesday, 29 March 2005 09:43:49 UTC