Re: Updated Conformance and Test Cases

On 5 Nov 2008, at 18:03, Alan Ruttenberg wrote:
[snip]
>>> A nit on the must accept/generate RDF/XML. Because of the syntactic
>>> restrictions in RDF/XML that prevent serialization of all RDF, for
>>> some OWL 2 Full document one MUST do something that isn't  
>>> possible. I
>>> suggest a footnote saying something about this. An alternative would
>>> be to specify that NTRIPLES must be acceptable as well.
>>
>> I would just amend the text to "MUST, if possible".
>
> I be happy with that but would prefer to couple it with something
> indicating that the other syntaxes SHOULD not be used in such a way as
> to not have it be possible for the RDF/XML to serialize their
> contents.

[snip]

But how else to do it? I mean, I think if you want to use URIs as  
properties that are incompatible with RDF/XML not only *should* you  
use one of the other serializations (OWL/XML i would  hope) but you  
*have* to.

You want a constraint on property names.

Given that the RDF documents say nothing in this regard, I don't  
think we should either. I don't think this "SHOULD" does anything  
more effective than the fact of RDF/XML (and our constraint on  
producing it and consuming it) already does.

Cheers,
Bijan.