pfps-22,pfps-23: "reserved names in abstract syntax", pushback

>From: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>
>
>I am still unhappy with the way that RDF sort of reserves certain things to
>itself, but also sort of allows anyone to do anything.

We've been asked [2] to be clearer about RDF namespace terms (full message 
below).

I'll try and pick out some key points for discussion.


1. "syntactic" URIs.

I think we could be stronger about these and say that they MUST NOT appear 
in an RDF abstract syntax graph.  (Does the syntax document already say 
this?)  Concepts (section 4) currently says that they "should not be used 
in RDF to denote any kind of resource" -- which might be strengthened to 
"must not".


2.  Status of rdfs:Class rdfs:Class rdfs:Class .

My view is that this is syntactically legitimate RDF, whose truth is 
neither asserted nor denied by the RDF specification.  Use of statements 
like this in RDF data is strongly discouraged, except as part of a semantic 
extension in conformance with the General Monotonicity Lemma [1].

Some extra words to this effect might be added to Concepts section 2.2.6:
   http://www.w3.org/2001/sw/RDFCore/TR/WD-rdf-concepts-20030117/#section-anyone

e.g.

[[
While legal RDF data may make assertions about concepts that are introduced 
by the RDF specification (the values, classes and properties indicated by 
specific RDF vocabulary terms), such use is strongly discouraged and may 
lead to RDF data that is useless when processed in accordance with normal 
RDF entailments.  As noted in the RDF formal semantics [RDF-SEMANTICS], 
semantic extensions may further constrain the meaning of RDF vocabulary 
terms in ways that conform to the General Monotonicity Lemma.
]]

Also, the paragraph in section 4 that currently reads:
[[
Vocabulary terms in the rdf: namespace are listed in section 5.1 of the RDF 
syntax specification [RDF-SYNTAX].  Some of these terms are defined by the 
RDF specifications to denote specific concepts.  Others have purely 
syntactic purpose (e.g. rdf:ID is part of the RDF/XML syntax) and should 
not be used in RDF to denote any kind of resource.
]]

could add something like:

[[
The meaning of non-syntactic RDF vocabulary terms is defined by the RDF 
Semantics specification [RDF-SEMANTICS].
]]


3.  what is the status of OWL's use of the RDF and RDFS vocabularies?

I regard these as fine as long as they conform with the General 
Monotonicity Lemma [1].  I can't easily tell, but I assume they do   I 
think this is covered above.

#g
--

[1] http://www.ihmc.us/users/phayes/RDF_Semant_Edit_Weak.html#MonSemExt

[2] http://lists.w3.org/Archives/Public/www-rdf-comments/2003JulSep/0006.html
     (message content below)


>Cc: www-rdf-comments@w3.org
>Subject: Re: [closed] pfps-22,pfps-23: "reserved names in abstract syntax"
>From: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>
>
>I am still unhappy with the way that RDF sort of reserves certain things to
>itself, but also sort of allows anyone to do anything.
>
>In particular, Section 2.2.6 says ``RDF is an open-world framework that
>allows anyone to make statements about any resource'' but Section 4 says
>``Certain URI references are reserved for use by RDF and should not be used
>in ways not supported by the RDF specficiations.''
>
>So, what is the status of, for example,
>
>         rdfs:Class rdfs:Class rdfs:Class .
>
>Is it a) completely unobjectionable, b) something that should not be done,
>or c) forbidden?  Section 2.2.6 argues for a); Section 4 argues for b) or
>maybe even c).
>
>Similarly, what is the status of OWL's use of the RDF and RDFS
>vocabularies?  (See
>http://www-db.research.bell-labs.com/user/pfps/owl/semantics/rdfs.html#5.2
>for the current editor's draft of the most relevant portion of the OWL
>specifications.)  Is this something that any formal specification can
>unobjectionably do, or is there something wrong with using the RDF and RDFS
>vocabularies in this fashion?
>
>The RDF Semantics document makes this even less clear as it explicitly
>mentions that semantic extensions may modify the meaning of rdfs:domain and
>rdfs:range (Section 4.1), but does not say anthing similar for most other
>elements of the RDF and RDFS vocabularies.
>
>peter
>
>
>
>From: Graham Klyne <gk@ninebynine.org>
>Subject: [closed] pfps-22,pfps-23: "reserved names in abstract syntax"
>Date: Tue, 01 Jul 2003 10:26:19 +0100
>
> > Peter,
> >
> > With reference to your comments raised in:
> > 
> http://lists.w3.org/Archives/Public/www-rdf-comments/2003JanMar/0148.html
> > and subsequent exchanges linked from:
> > 
> http://lists.w3.org/Archives/Public/www-rdf-comments/2003JanMar/0314.html
> > specifically with reference to the issue of reserved names in the RDF
> > syntax, and the notion of uses "sanctioned by" RDF, which were 
> crystalized in:
> > 
> http://lists.w3.org/Archives/Public/www-rdf-comments/2003JanMar/0194.html
> >
> > The RDFcore working group has resolved per:
> >    http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2003Apr/0207.html
> > (agendum 16) to revise the text along the lines of:
> >    http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2003Apr/0201.html
> >
> > Revised text can be previewed in the editors' working draft at:
> >    http://www.w3.org/2001/sw/RDFCore/TR/WD-rdf-concepts-20030117
> >
> > Could you please respond, copying www-rdf-comments@w3.org, indicating
> > whether or not you regard your comments have been adequately addressed.
> >
> > Thank you for your attention,
> >
> > Graham Klyne
> > (for RDFcore working group)
> >
> > #g
> >
> >
> > -------------------
> > Graham Klyne
> > <GK@NineByNine.org>
> > PGP: 0FAA 69FF C083 000B A2E9  A131 01B9 1C7A DBCA CB5E

-------------------
Graham Klyne
<GK@NineByNine.org>
PGP: 0FAA 69FF C083 000B A2E9  A131 01B9 1C7A DBCA CB5E

Received on Tuesday, 1 July 2003 12:02:27 UTC