- From: <Patrick.Stickler@nokia.com>
- Date: Thu, 9 Sep 2004 11:47:26 +0300
- To: <timbl@w3.org>, <www-tag@w3.org>
> -----Original Message----- > From: www-tag-request@w3.org > [mailto:www-tag-request@w3.org]On Behalf Of > ext Tim Berners-Lee > Sent: 08 September, 2004 20:13 > To: 'www-tag@w3.org' > Subject: Re: Information resources? > > > > In http://lists.w3.org/Archives/Public/www-tag/2004Sep/0028.html Tim > Bray said, > > > > On Sep 7, 2004, at 7:17 AM, Norman Walsh wrote: > > > > > The notion of "resources" and "information resources" is, from my > > > perspective, a compromise designed to allow two world views to > > achieve > > > consensus. > > > > I think the distinction is useful and worth talking about > in webarch, > > because it's something that people feel is real. Obviously the > > condition of being an information resource is something you can only > > test by attempting to dereference, but there's clearly a distinction > > between something that you normally expect to be there and > if it's not > > you assume breakage, and something that you've never seen and if you > > get a 404 you say "Oh well, just a name I guess". So at some level, > > being an information resource is a *social* condition, a matter of > > expectations and behaviors. > > This is not the definition of an Information Resource I use. > > Information resources are the sort fo things which convey > information. > You can bicker about the edges but... > > here are some things which are information resources > > A picture > A poem > An email message > A document > A web page > A movie > A mail message > The book "Moby Dick" > The Declaration of Independence of the USA > > Here are some things which I do not think of as information resources: > > A dog > A person > The concept of parenthood > The USA > The W3C > Cornflower blue > > Here are some properties which some IRs typically have > > Subject (although this may not be well defined) > Creator > Creation time > Translation into French > Digital encoding in PNG > > Information Resources are a major part of the Web > architecture that not > to have a word for them. > > The fact that a given information resource, identified by some URI, > may at some times be accessible and at other times not, > doesn't change > the nature of the information resource. > > The HTTP spec didn't need to distinguish between dogs and pictures of > dogs, as HTTP doesn't deal in the difference. Humans can cope with > ambiguity of this sort. But the architecture document has to, in my > opinion, be able to distinguish between them. But does it really? Is this not rather the job of the semantic web? I.e. why can't the web architecture be "agnostic" about the inherent nature of resources? A URI denotes a resource. A "web resource" has web-accessible representations. The web machinery facilitates interaction with representations. That's it, really. Then the semantic web steps in and facilitates descriptions of those resources, and one can say that some resource is an "information resource", or a "web page" or a "poem" (or a "dog" or a "color" etc.). The web architecture itself does not have to care about what those resources are -- only about whether any representations are associated with those resources and how to facilitate interaction with those representations. Yes, it is quite true that resources corresponding to "bodies of information" are central to the web -- because *people* use the web to exchange information -- but the nature of those resources as bodies of information is *not* central to the machinery of the web; and thus no distinction or reference to "information resources" is necessary in AWWW. Regards, Patrick resources as "bodies
Received on Thursday, 9 September 2004 08:47:45 UTC