W3C home > Mailing lists > Public > www-tag@w3.org > February 2009

Re: Introduction to URIs (was RE: WebArch introduction, sort of)

From: Tim Berners-Lee <timbl@w3.org>
Date: Sun, 1 Feb 2009 06:53:06 -0800
Cc: "'Alan Ruttenberg'" <alanruttenberg@gmail.com>, "'Booth, David \(HP Software - Boston\)'" <dbooth@hp.com>, "'Henry S. Thompson'" <ht@inf.ed.ac.uk>, <www-tag@w3.org>
Message-Id: <B12A5BCF-3523-4353-A518-695E208283FF@w3.org>
To: Larry Masinter <LMM@acm.org>

On 2009-01 -26, at 18:56, Larry Masinter wrote:

> I might ask "what is correct?" and "what is a non-information  
> resource"?
> My point wasn't to debate the merits of one definition or another,
> it was exactly to point out that the question rests entirely on
> terms that have no good definition by themselves, and that you might
> as well resolve them by definition rather than continuing to debate
> them.

The conventional wisdom that one should first define one's terms and  
write carefully using them does not apply.
We are defining a system from new. (Philosophical engineering).
We define new concepts, and largely in terms of each other.
We will need terms, and the design ties them together and to a certain
extent in places to outside world and other well-defined things, but  
to a limited
extent only.  Much of the meaning of a term is involved in its
relationship to other terms, which seems circular and unhelpful to the  

There have been years of discussion of the right english words
to define an Information Resource.  I tend to connect it ti the word
"document" but all words like that have different associations for  

So if you look at what you can do with a URI

- You can expect to get the same information next time (typically)

- You can get different content-type of the same thing

There are important conventions:

- Some change with time, some don't

and there are social aspects:

- The owner of the domain name has the ability to control what you get

These all are examples of properties of information resources which  
follow mainly from the protocols.


In fact Information Resource is in some ways a better defined term
than 'document' , but it is defined by what you can so with it.
So yes, taking an operational approach is necessary but you
have to do it at the semantic web level.
At the HTTP level there are no tests you can do which allow one to tell
whether http://wikipedia.org/page/Paris (or whatever it is -- I am on  
a plane)
denotes Paris or a web page. But
semantic web systems are very aware of the difference
whether http://dbpedia.org/resource/Paris  is a town or a web page,
but semantic web systems are very aware of the difference in practice.
So semantic web systems have to pick what to use an hashless http URIs  
denote, and we picked the document not the town.

> If it isn't clear, I'm in favor of taking an operational approach.

If that means what I just said then I'm all in favor. :)

This is, by the way, very very very well-trodden ground.

Received on Sunday, 1 February 2009 20:25:37 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:33:00 UTC