Re: Should tags be URNs? (was Re: Proposal: 'tag' URIs)

> I would say that, like any 'name' or 'identifier', URIs aren't 
> intrinsically identifiers: they're intrinsically strings that are unique in 
> some context; it's how they are used that make them identifiers.

Of course -- that is the definition of "identifier".

> I also want to say that anything that can be usefully expressed in our 
> domain can be specified in terms of sets and functions (alright, alright: 
> categories). I would like to rid us of 'abstract' and 'conceptual'  and 
> replace them with simple mathematical objects that we can all understand, 
> reason about and implement inside machines. I'll try and do that, below.

No, you are missing the point.  The terms abstract and conceptual are every
bit as formal as any mathematical definition you might invent, and far more
likely to serve the intended purpose.

What is a URI?  It is an identifier for all those things an author might
wish to identify as the target of a relationship.  It is from that goal that
we obtained the current definition of URI and its use within HTTP and
various data formats.  Inventing mathematical definitions that are
more restrictive than the one found in my dissertation will serve only
to define something more restrictive than what we intended to implement,
and thus not useful as a substitute in any further formalisms.

In other words, live with the ambiguity and learn its advantages.

   http://www.ics.uci.edu/~fielding/pubs/fielding_diss.pdf

Cheers,

Roy T. Fielding, Chief Scientist, eBuilt, Inc.
                 2652 McGaw Avenue
                 Irvine, CA 92614-5840  fax:+1.949.609.0001
                 (fielding@ebuilt.com)  <http://www.eBuilt.com>

                 Chairman, The Apache Software Foundation
                 (fielding@apache.org)  <http://www.apache.org/>

Received on Friday, 4 May 2001 16:39:51 UTC