W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > November 2002

Re: XMLLiteral belongs in RDF namespace, not RDFs

From: Graham Klyne <Graham.Klyne@MIMEsweeper.com>
Date: Sun, 10 Nov 2002 11:36:03 +0000
Message-Id: <>
To: pat hayes <phayes@ai.uwf.edu>
Cc: w3c-rdfcore-wg@w3.org

Short response:

I think we're moving toward using rdf:... rather than rdfs:..., and I'm 
happy with that.


At 02:43 PM 11/9/02 -0600, pat hayes wrote:
>>Oh dear, I haven't being paying attention properly.  Apologies. Something 
>>was niggling me about this, but I now think I see what it was:
>>The datatyping extension is both syntactic and semantic.  As a syntactic 
>>extension, it naturally belongs with the core language
>It needs to use rdfs:Datatype to signal a recognized dtype, that's 
>essentially syntactic.


>>.  But as a semantic extension, it fits more comfortably (IMO) with the 
>>schema material.  Hmmm... I don't know what to suggest as a solution.
>Wait.... .solution? What exactly is the problem here?

The "problem" I perceive, partly in response to Dave's comment, is that 
datatyping is core language syntax, but that the corresponding semantics of 
datatype-entailment don't really sit there.  I don't think there's any 
fundamental technical problem, but a challenge of presentation that the 
relevant material is presented cleanly (e.g. avoiding dependencies like, 
for example, the RDF core specification on RDF schema).

>>  (DanC's approach, which we turned down, starts to look more attractive.)
>>That's not helpful... thinks...  the only thing I can think of that seems 
>>reasonably coherent is to bring the datatype URI into the core (rdf:), 
>>even if it is "adding a new term that has a meaning".  It would not be 
>>the first such term in the core language;  e.g. we already rdf:type, 
>>which has some defined semantics in an RDF-interpretation.
>I don't follow you here. Did you mean to say in an RDFS-interpretation? 
>Because that wouldnt really be accurate.

No, I meant RDF-interpretation.  I was responding to Dave's comment:
Since it is adding a new term that has a meaning (will get some
description in an RDF schema document) rather than something for
building the RDF/XML syntax, I thought our policy was to add stick
them in RDFS namespace.
and suggesting that adding additional semantic rules in the RDF namespace 
would not be overturning some established policy.

>>Is there any reason why a datatyped-interpretation has also to be an 
>No deep reason. Its tricky to say much without mentioning rdfs:Dataype and 
>rdfs:Literal , is all.

Yes... the challenge of presentation?

>I think we discussed having rdf:Datatype and rejected it for some W3C 
>procedural reason (??).

I don't know now.

I think we're moving toward having rdf:Datatype rather than rdfs:..., and 
I'm happy with that.

(For now, at least.  Having read through the new stuff on datatyping, and 
thinking of Aaron's comments about complexity, I do find DanC's approach to 
datatyping more appealing (sans the lexical form constraint), by virtue of 
avoiding a whole slew of new mechanism.  But, this isn't the time to 
question group consensus and I'm not about to rock the boat here.  But I 
think it's something we might wish to contemplate if there's community 
concern about datatyping in the last call period.)


>>At 05:53 PM 11/8/02 +0000, Dave Beckett wrote:
>>>  >>>Dan Connolly said:
>>>>  I see
>>>>    rdfs:XMLLiteral
>>>>  in
>>>>  http://www.coginst.uwf.edu/~phayes/RDF%20Model%20Theory_Oct_draft.html
>>>>  That should be rdf:XMLLiteral, right Dave?
>>>No, we agreed rdfs:XMLLiteral
>>>I noted this danger earlier this week.
>>>>  eek... it's there in the syntax editor's draft
>>>>  too:
>>>>  If literal-language is the empty string then the value is the
>>>>  concatenation of """ (1 double quote), the value of the literal-value
>>>>  accessor and ""^^<http://www.w3.org/2000/01/rdf-schema#XMLLiteral>" (1
>>>>  double quote).
>>>>  -- http://ilrt.org/discovery/2001/07/rdf-syntax-grammar/
>>>It is in the soon to published WD too.
>>>>  Let's please be careful... there is no
>>>>  dependency on RDFS from RDF.
>>>Since it is adding a new term that has a meaning (will get some
>>>description in an RDF schema document) rather than something for
>>>building the RDF/XML syntax, I thought our policy was to add stick
>>>them in RDFS namespace.
>>>>  I thought we could get away with a combined
>>>>  model theory spec, at least for a while.
>>>>  But I think that time is ending.
>>>>  And I'm starting to wonder about the primer...
>>>>  ah; the primer is cited non-normatively
>>>>  from that syntax draft; as long as we
>>>>  do that, it can have both RDF and RDFS.
>>Graham Klyne
>IHMC                                    (850)434 8903   home
>40 South Alcaniz St.                    (850)202 4416   office
>Pensacola                               (850)202 4440   fax
>FL 32501                                        (850)291 0667    cell
>phayes@ai.uwf.edu                 http://www.coginst.uwf.edu/~phayes
>s.pam@ai.uwf.edu   for spam

Graham Klyne
Received on Sunday, 10 November 2002 06:47:38 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:24:18 UTC