- From: Harry Halpin <hhalpin@ibiblio.org>
- Date: Fri, 28 Sep 2007 10:54:06 -0400
- To: "Williams, Stuart (HP Labs, Bristol)" <skw@hp.com>
- Cc: Pat Hayes <phayes@ihmc.us>, Dan Connolly <connolly@w3.org>, Tim Berners-Lee <timbl@w3.org>, Technical Architecture Group WG <www-tag@w3.org>, Susie Stephens <susie.stephens@gmail.com>
Williams, Stuart (HP Labs, Bristol) wrote: [snip] >> But the key point is that the >> distinction has to do with whether it can be >> *accessed* by transfer protocols, rather than with the way it >> can be encoded or represented. The essential characteristics >> of a billboard can be conveyed in an image, but if the image >> has never been digitized then its not an information resource >> (yet). The problem is simple and practical: Given any "resource/thing" that one wishes to talk about on the Web (i.e. using RDF, putting up web-pages, etc.), is given resource X an information resource or not? That's a question with real ramifications - i.e. should one use a 303 redirect/# convention or not! > FWIW, the wording of the TAG/webarch defn of 'information resource' > arose during our meeting in Basel [1]. > > Earlier attempts to offer the term "Web resource" for things that > respond with 200 and a representation [2] failed on ground which now > escape me. > > The thing that I recall about the "essentail characteristics being > conveyable in a message" definition was that it was about the > *potential* of such conveyance rather that the actuality of it. The > notion was, IIRC, intended to be about some Shannoesque notion of > information content rather than the fact or otherwise of a 200 response > when using a particular transfer protocol. > > I mention this here because there seems to be a tendency in these > threads to offer a definition whereby "information resources" are just > those which in fact respond to an http request with a 200 response code > - whereas the TAG definition was intended to be broader - anything whose > essential nature is information. > This is precisely the problem. Because the notion of "information" is historically underdefined (Shannon himself said it had only to do with encoding, not with denotation or semantics), given a URI for X whether or not it is an "information resource" or not is a matter of philosophical debate - i.e. not a cut and dry question for your average SemWeb hacker on the streets. > So, wrt to the billboard poster that you mention above, the image on the > poster would 'fit' the definition whereas the poster itself might not - > having a mass, being made of some material that will deteriorate over a > few weeks; > I think the only logical answer is to junk the underdefined notion of "information" resource and replace/redefine it by something similar to "Web resource" - saying that an information resource is a resource that is composed of the total of all representations it serves (using HTTP 2xx) over time. Anything else is a non-information resource. The "over time" deals with the vagarities of time, and the "total of all representations" deals with the possibility of conneg serving different representations. So, if what you want to mint a URI X, all you have to say is that "what I want URI X to mean is totally composed by all representations hosted there or not?" That answer can be given by the average hacker on the street. So, TimBL's web-page would be an information resource, TimBL would not. A URI that gave a representation (like a .JPG via 200) of the image given by a billboard message would be one, but under this definition, the example given by Stuart, a URI for an image given by a billboard that *does not actually host a representation of that image* would *not* be an information resource, and so would need a 303 or #. I think this alternative makes a lot of sense, and would like to hear why it does not. I admit the more abstract idea of "information resource" is coherent, but I think for people wanting to actually start deploying Semantic Web URIs and who are in the dark about this 303 business, we need a more practical definition like the one I gave above. > Stuart > -- > [1] http://www.w3.org/2001/tag/2004/10/05-07-tag#infores234 > [2] > http://lists.w3.org/Archives/Public/public-webarch-comments/2004JulSep/0 > 086.html > > <snip/> > > >> Pat >> >> >>> -- >>> Dan Connolly, W3C http://www.w3.org/People/Connolly/ >>> > > >> -- >> --------------------------------------------------------------------- >> IHMC (850)434 8903 or (650)494 3973 home >> 40 South Alcaniz St. (850)202 4416 office >> Pensacola (850)202 4440 fax >> FL 32502 (850)291 0667 cell >> phayesAT-SIGNihmc.us http://www.ihmc.us/users/phayes >> > > Stuart > -- > Hewlett-Packard Limited registered Office: Cain Road, Bracknell, Berks > RG12 1HN > Registered No: 690597 England > > > -- -harry Harry Halpin, University of Edinburgh http://www.ibiblio.org/hhalpin 6B522426
Received on Friday, 28 September 2007 14:55:38 UTC