- From: Graham Klyne <Graham.Klyne@MIMEsweeper.com>
- Date: Fri, 01 Nov 2002 18:17:09 +0000
- To: Brian McBride <bwm@hplb.hpl.hp.com>
- Cc: pat hayes <phayes@ai.uwf.edu>, w3c-rdfcore-wg@w3.org
At 09:51 AM 11/1/02 +0000, Brian McBride wrote:
>At 23:46 31/10/2002 -0600, pat hayes wrote:
>
>>Just how minimal do we want the list semantics to be? In particular, is
>>this satisfiable? :
>>
>>7.
>>rdf:nil rdf:rest _:xxx .
>>
>>? Or can I rule that out? If not, our claim that lists are bounded seems
>>rather hollow, and that was the point of having them in the first place.....
>
>Here I want to float an idea I have mentioned offline to Pat.
>
>Would it make sense to restrict the structure of collections in the
>*abtract syntax*. Don't worry Dave, I don't think it affects the XML
>syntax - it can only produce well formed lists already. We write the
>abstract syntax so that lists must be syntactically well formed. Anything
>else is not well formed RDF.
I like the thinking, but...
This might open the possibility that the merge of two well-formed graphs is
itself not well-formed.
Example:
Graph 1:
ex:head rdf:first ex:item1 .
ex:head rdf:rest rdf:Nil . <<<*** is that rdf:Nil or rdf:nil, BTW ???
Graph 2:
ex:head rdf:first ex:item2 .
ex:head rdf:rest rdf:Nil .
Separately, these are well-formed, but if merged then ex:head then has two
rdf:first properties. Is this a problem?
#g
-------------------
Graham Klyne
<GK@NineByNine.org>
Received on Friday, 1 November 2002 14:14:15 UTC