RE: Profiling in the News: Was: RE: Taxonomies, URN's etc..

["Bullard, Claude L (Len)" <clbullar@ingr.com>]

Mapping is the *process* whereby the URI becomes 
the identifier.  An SSN is not a URI.  I can 
use a URI in a namespace for disambiguating 
names.   No problem.  I can also write the 
schemas such that names are never ambiguous. 

I didn't rule out using URIs.  I said I don't 
need them to build schemas or declare property 
sets.  I need them when I want to use those 
reliably on the Internet.   Some systems never 
use the Internet but can use XML.  HumanML 
can be applied to non-Internet systems.

We can apply RDF and or XML Schemas or even 
RELAX NG but we have to assess the buy in 
given customers who are thralls to tools.  Today 
we can go faster and get on more systems using 
the schemas.  The question is technically one 
of representational power but in project time, 
one of isolating the domains.  For that reason, 
I welcome the parallel examples.  For the same 
reason, I don't like to see HumanML labeled as 
a Semantic Web project.   Too many contracts 
and not enough working lawyers. 

We should be assessing the prototype schemas, 
not debating URIs, RDF, the Semantic Web or 
other initiatives.  If we can get the property 
sets scoped, defined and documented, we have 
time to work out alternative representations 
for different processing environments.

Len 
http://www.mp3.com/LenBullard

Ekam sat.h, Vipraah bahudhaa vadanti.
Daamyata. Datta. Dayadhvam.h


-----Original Message-----
From: Sean B. Palmer [mailto:sean@mysterylights.com]

The process of identifying resources on the Semantic Web is made easier by
using URI references, but it is also useful to talk about some resource as
an existentially quantified node using a set of proerties that relate to
it... some of which may be unambiguous, either individually, or when used
as a combination of properties.