W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > April 2003

Re: help needed with RDFS issue pfps-12, wellformedness of lists

From: pat hayes <phayes@ai.uwf.edu>
Date: Sat, 26 Apr 2003 14:26:56 -0500
Message-Id: <p0521060cbad07be8d84f@[]>
To: Dan Brickley <danbri@w3.org>
Cc: w3c-rdfcore-wg@w3.org

>The RDF Schema document
>    http://www.w3.org/TR/rdf-schema/#ch_collectionvocab
>describes lists as though they were always "well formed", which they are not.
>I can think of two ways of tackling this issue. Either a 'health 
>warning' approach,
>where we note that partial or broken rdf:List descriptions are possible,

That is what we have at present, and I think we should leave it at 
that.  If we try to put list-well-formedness into the RDF syntax we 
will never get out of the resulting tar-pit, and Peter himself 
requested that we remove it from the model theory. To resolve the 
issue, do a small editorial tweak to the wording in the document so 
it doesn't make the assumption.

>or by
>trying to account for the rules for being a wellformed rdf:List.

But if you want to explore this, we did the relevant work already in 
an email to Brian a while ago.  I think its in the archives somewhere.

There is a third option, which is to not impose a syntactic 
constraint on RDF graphs, but to impose a semantic constraint that 
all 'list-well-formed' RDF graphs must actually denote proper lists. 
I already floated this idea and did the model theory for it, but the 
WG decided to remove it in part because Peter himself requested that 
we remove it: not because it was harmful to OWL (in fact it would be 
marginally useful to OWL) but on aesthetic grounds of RDF having a 
'minimal' semantics. I still think this was a bad decision, but I 
don't see any point in re-opening it at this stage.

>The latter was
>begun in the Peter/Brian dialog cited above:
>>  > A rdf:List is well formed if it meets either of the following conditions:
>>  >
>>  >    o it is rdf:nil
>>  >    o - it has exactly one rdf:first property,
>>  >      - and it has one rdf:rest property
>>  >      - and the value of its rdf:rest property is a well formed list.
>>This is not sufficient to describe well-formed lists!  (Think of infinite
>>or circular lists.  Also think of what happens if rdf:nil is the subject of
>>a triple whose predicate is rdf:first or rdf:rest.)
>Just so.
>Right, I think we've got the point where we have clarified what the issue
>is, but we are probably going to have to think a little more about how best
>to address it.
>I need help in progressing this towards an issue closure proposal. 
>Could someone
>take a crack at refining the above exchange into a more solid
>characterisation of rdf:List? Especially re the circular and infinite concerns
>Peter raises.
>Perhaps we should also write, "An RDF description of a rdf:List is 
>well formed"
>rather than "A rdf:List is well formed..."?

Yes, its the graph that is 'wellformed'; but use this phrase in scare 
quotes in this context, please, since ill-formed-list graphs are 
well-formed RDF. . See 

IHMC					(850)434 8903 or (650)494 3973   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
Received on Saturday, 26 April 2003 15:27:02 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:24:22 UTC