Justification for new node type

	
I would like the proponents of the recent proposal for
a new datatyped literal node type to justify why URIs
cannot be used.

In essence, what is being proposed, as I understand it
is that we would have a new labeled graph node time, where
the label would include the datatype and the lexical
form, which would globally, unambiguously, and consistently
denote a particular datatype value.

But RDF already has a type of labeled node that is 
intended to be used to globally, unambiguously, and
consistently denote a resource. It's a URIref node.

Just as RDF need not look inside URIrefs to test
potential denotational equality of string-unequal URIs,
likewise RDF will not itself look at datatypes and
make any determination about equality of datatyped
literals which have string-unequal lexical forms
with identical datatypes.

In both cases, applications are free to parse the URIs
themselves to make such determinations, but both kinds
of comparisons are external to RDF itself. Therefore,
there is no reason why a URI would be any more restrictive
to RDF level processing than some other new form of
node label -- which will be treated just as opaquely
as URIs.

Therefore...

Other than the (political) argument about registering
a new URI scheme to denote datatyped values, such as
http://ietf.org/internet-drafts/draft-pstickler-val-01.txt,
I would like to hear technical (not political) justification
for why URIs cannot be used as the compact local
datatyping idiom.

I believe that there is a burden on the proponents of
this recent proposal to clearly demonstrate (to the
agreement of the WG) that URIrefs do not do the job
as a compact, single node denotation of datatype values.

Patrick

--
               
Patrick Stickler              Phone: +358 50 483 9453
Senior Research Scientist     Fax:   +358 7180 35409
Nokia Research Center         Email: patrick.stickler@nokia.com

Received on Friday, 9 August 2002 12:05:29 UTC