RE: URI for language identifiers

> -----Original Message-----
> From: ext Miles Sabin [mailto:miles@milessabin.com]
> Sent: 01 April, 2003 16:49
> To: www-rdf-interest@w3.org
> Subject: Re: URI for language identifiers
> 
> 
> 
> Patrick Stickler wrote,
> > Either you have not understood what I was saying, or you are simply
> > wrong (or both ;-)
> >
> > The denotation of a given URI is that which the owner of that URI
> > specifies. Period.
> 
> Umm ... likewise ;-)
> 
> There's no such thing as "the owner of that URI". Period.
> 
> The most you can say is that there's an owner of the DNS 
> name. But for 
> complete URIs there's no spec, no registration procedure and no 
> possibility of legal redress which would prevent someone choosing to 
> associate any meaning whatsoever with an arbitrary URI _without_ 
> agreement from the owner of the embedded DNS name.

I think that if I defined a public vocabulary grounded in http://www.w3.org/
that the W3C would send a team of lawyers to my door.

If I started adding all sorts of new properties to RDF and OWL, grounded in
the RDF and OWL namespaces, I think that a very large, if not unanimous,
view would be that I was invading space that does not belong to me -- i.e. that
I have no authority or right to posit such properties using URIs grounded in
a web authority I don't own.

While it may be true that I cannot point to a specific line of a specific spec
that clearly defines such a concept of URI ownership, I think it's pretty
naive to presume that it doesn't exist.

> Of course, the owner of the DNS name can control the effect of public 
> network operations on such URIs, because she controls the public name 
> to address mapping (which is exactly what ownership of a DNS name 
> amounts to) and presumably also controls the network software 
> accessible at the resolved address.  But this isn't 
> necessarily relevant 
> to an RDF application, so not necessarily an obstacle to someone who 
> wants to associate a local meaning with the URI for RDF related 
> purposes.

Exactly.

The owner of the DNS name could *only* control access to representations,
but *NOT* usage of those "rogue" URIs on the semantic web, since RDF has
no concept of representation or dereferencing and no need for there to
ever be anything obtainable by a given URI. And in the case of vocabulary
terms, it is the exception than the rule that anyone would attempt to
dereference such a URI.

But just because use of a given URI can happen in RDF-space irrespective
of control over network operations does not mean that the absence of such
control equates to an absence of ownership.

> Feel free to deplore this status quo ... but you can't change 
> it simply 
> by asserting the contrary.

Nor can you say that ownership of URIs does not exist as a clear and
widely accepted concept simply because the specs do not define it
explicitly.

There are two status quo's it seems. That of the specs and that of
the real world. Sometimes the two are in sync. Sometimes they are not.
As far as I'm concerned, silence on the part of one does in no way 
invalidate the other.

Patrick

Received on Wednesday, 2 April 2003 01:41:39 UTC