Re: what is this? was: Re: now://example.org/car

Dan Connolly wrote:

> On Wed, 2002-10-09 at 06:51, Jonathan Borden wrote:
...
> > Two people might effectively communicate without _directly_ speaking of
web
> > pages. What I am saying is that URIs are _words_ and words are used to
mean
> > what the people use them to mean.
> >
> > The really cool thing about HTTP words is that you click on them to find
out
> > what they mean -- but still, they are interpreted in context just as
other
> > words do.
> >
> > For example:
> >
> > [[
> > Hey Bill, do you like the layout of
> > http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=1772811392
> > Should we change the fonts?
> > ]]
> > vs.
> > [[
> > Hey Bill, is
http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=1772811392
> > your very favorite car?
> > ]]
> >
> > Same representation, but each use of the word has a different meaning --
> > depending on context.
>
> Yes, people are sloppy.
>
> But are you suggesting that this sort of ambiguity is acceptable
> in formal languages, such as technical specifications
> of XML formats? I hope not.
>

Hardly.

I am suggesting that we can easily assert:

http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=1772811392 :favoriteCarOf
:Bill .

or perhaps

http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=1772811392 :usesFont
:Arial .

and that given a sufficient ontology, we might conclude that this URI is
being used in a _contradictory_ fashion, that something which has the
property "usesFont" cannot also have the property "favoriteCarOf".

What prevents me from asserting whatever I please about any given URI -- as
long as the assertions are consistent?
What necessitates that any particular URI have a specific rdf:range on its
rdf:type? (to speak in a less human but more formal language). Doesn't
RDF(S), for example, treat URIs as essentially opaque tokens ? (to consider
formal systems)

Jonathan