W3C home > Mailing lists > Public > www-tag@w3.org > September 2002

Re: My action item on Moby Dec, issue 14, etc

From: Tim Bray <tbray@textuality.com>
Date: Fri, 20 Sep 2002 09:38:47 -0700
Message-ID: <3D8B4F17.3050902@textuality.com>
To: www-tag@w3.org

Michael Mealling wrote:

> Hmm... I don't see a difference between these two viewpoints. The system
> is only made up of two things: a URI and a Resource. A URI can't exist
> without a Resource and a Resource can't exist without a URI, and the
> mapping is 1:1 exclusive. Within that system it seems pretty unambiguos.
> But it all comes back to the fact that some prefer to use the 
> term 'resource' to mean a 'a bag of bits' instead of 'the thing
> that exists when a URI is created'. The first definition of 'resource'
> exists without any context involving URIs so there is no defined mapping
> between a URI and that definition, hence it is very ambiguous. 
> I think what Tim Bray is saying is that attempting to use definitions for 
> 'resource' other than 'the thing that is created with a URI is created'
> is dangerous and not within the realm of provability or usefullness
> at the architecture level.

I think Micheal has hit it on the head.

There are only two ways you can have it:

(a) a resource is defined as that which a URI identifies (I think this 
is consistent with 2396, if you believe "has identity" is a synonym for 
"you could write a URI").  In this case, the mapping is 1:1 by 
definition and our so-called principle #2 is an entirely vacuous tautology.

(b) a resource is something separate and real that is labeled with the 
URI.  If this is the case, there is no way to be sure that the domain 
name hasn't changed ownership or example.com, Inc. hasn't reorganized 
and thus the URI now points to a *different* something.  I would point 
out that we have another principle about persistence which states 
(correctly) that you shouldn't do this.  In which case our so-called 
principle #2 is *wrong*.

Let me re-iterate that last point. Right now Webarch says
  #2. URIs are unambiguous
  #7. URIs should be persistent

If #2 is really true, we shouldn't have to say #7.  I quote from Webarch 
section 2.3 "On the other hand, if tomorrow, the same URI identified a 
different resource ... ".

I would be delighted to shown a consistent worldview that's not one of 
(a) or (b) and makes principle #2 either meaningful or true.

This requires more thought, but I'm not 100% sure that Webarch has to 
come down on the side of (a) or (b).  Choice (a) is easier to write 
about and doesn't break anything.  I lean to (b) because I think that 
resources *are* real, otherwise the Web wouldn't be interesting.  -Tim
Received on Friday, 20 September 2002 12:38:55 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:55:54 UTC