RE: web proper names

> Yes, a resource can be anything, thus a URI can identify 
> *anything*, but
> it would be useful to be able to distinguish explicitly between these 
> two types of things (for example, in our previous example, 
> while a human
> could read the comments in a RDF Schema or the Dublin Core 
> documentation
> to know the correct use of "creator", a machine could not). 

Simply publish some RDF statements about the resources in
question so that machines (and humans) would then be provided
with explicit information about those resources rather than
having to guess based on either the particular behavior of
a web server or the lexical properties of the URI in question.

> Our proposal builds without any problems on current web use and is not
> dependent on RDF, the Semantic Web in general, or the TAG.

IMO, the proposed wpn: URI scheme is in direct conflict with
the principle of URI opacity in that it promotes the assumption
of certain semantics based on the lexical properties of the URI
itself.

It also, as I've pointed out before, limits the web accessibility
of the resource thus denoted, unless one introduces additional
(unnecessary) machinery into the mix.

1. Simply use http: URIs to denote your resources.
2. Be sure to denote one and only one resource with a given URI.
3. Make representations of the resource available via that URI.
4. Make statements in RDF to explicitly describe the resource.

There is no real technical problem addressed by the wpn: URI scheme,
thus, such a "solution" is not needed.

The social/behavioral problem of what a given URI denotes
and how different resources are related is fully addressed
by the presently deployed web and semantic web technology.

What is needed are well defined, commonly employed ontologies
to capture the information about these resources that folks
seem to care so much about. The infrastructure itself is IMO
ready and waiting...

Regards,

Patrick

Received on Tuesday, 21 September 2004 09:15:05 UTC