Re: Attention Users! (RDF Core WG Decisions)

On Monday, June 25, 2001, at 03:21  PM, Karsten-A. Otto wrote:

> With the solution of rdfms-empty-property-elements the meaning 
> of <rdf:li/>
> is clear, but though it may be usefull it is not allowed by the 
> current M+S
> production rules. The same applies to <rdf:RDF/> (an empty 
> model). Please add
> some rules for these cases to the production set, otherwise RDF is not
> compatible to XML which defines <ns:something/> as an abbreviation of
> <ns:something></ns:something>.

Thank you for your comment. Yes, the Working Group is well aware 
of this and hopes to choose a system of production rules that 
does not make this mistake in the next revision of the spec. 
There seems to be consensus that for the current grammar, the 
abbreviation form is equivalent to the extended form. This will, 
of course, be fixed in the future.

> The example you give in the solution to 
> rdf-containers-formalmodel is not
> valid RDF according to the current M+S production rules. They 
> only allow
> the rdf:_n form as attributes, for elements the rdf:li form 
> must be used.
> Accordingly, current parsers transform rdf:li XML to rdf:_n 
> triples, and
> serializers transform rdf:_n triples back to rdf:li XML. If the 
> example is
> valid, I assume you are going to allow rdf:_n XML elements in 
> the transfer
> syntax. But then please explain when to use each form, and when 
> to transform
> between them.

This is part of an upcoming decision about the 
rdf-container-syntax issues. It seems likely that the group will 
agree that the use of both rdf:li and rdf:_n are equally valid. 
A full writeup of this decision is forthcoming.

Personally, I would recommend use of the rdf:_n syntax always, 
as this is much simpler.

--
       "Aaron Swartz"      |              The Semantic Web
  <mailto:me@aaronsw.com>  |  <http://logicerror.com/semanticWeb-long>
<http://www.aaronsw.com/> |        i'm working to make it happen

Received on Monday, 25 June 2001 20:50:29 UTC