[doap-interest] [ANN] doap:store, a collaborative DOAP projects directory

[Christopher Schmidt]

On Mon, Dec 18, 2006 at 03:21:29PM +0000, Uldis Bojars wrote:
> Hi Chris,
> 
> On 12/18/06, Christopher Schmidt <crschmidt at crschmidt.net> wrote:
> >On Mon, Dec 18, 2006 at 01:33:47PM +0000, Uldis Bojars wrote:
> >>
> >> OK, I see what is the problem.
> >>
> >> DOAP is based on RDF document format. Please check your document in
> >> the RDF validator - it is not a valid RDF document.
> >> http://www.w3.org/RDF/Validator/
> >
> >The validator is wrong. It's heuristic for checking whether content is
> >RDF is failing, but the content itself is still valid RDF.
> 
> Thanks for correcting me. Indeed, RDF/XML spec. says: "When there is
> only one top-level node element inside rdf:RDF, the rdf:RDF can be
> omitted although any XML namespaces must still be declared."
> 
> Still, there is a possibility that PingTheSemanticWeb expects RDF/XML
> sources to have the top-level rdf:RDF. Seems it can be omitted as
> well, but still it can be a reason why PingTheSemanticWeb would not
> accept this DOAP profile.

Yep.

> >> Try adding:
> >> <?xml version="1.0"?>
> >> <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
> >> to the front of the document and </rdf:RDF at the end.
> >
> >DOAP does not allow this. Adding this to a DOAP document makes it
> >non-DOAP.
> 
> Do you want to say that DOAP is not RDF?
> If it were RDF then wrapping it in "<rdf:RDF>" should not change its 
> validity.

That's exactly what I'm saying, yes. DOAP is a specific subset of
RDF/XML, deisgned to be *easy for XML parser to understand*, but also
designed to fit well into the RDF/XML framework that makes itm ore
generally useful. 

> Noticed that you use the "<rdf:RDF> part in the DOAP profile of DOAPer 
> anyway:
> http://crschmidt.net/doap/doaper.rdf

That's correct, and it's incorrect DOAP. (It is correct RDF.)

> I find it peculiar if a valid DOAP profile becomes invalid if parsed
> from RDF/XML and serialised again in RDF/XML (which most probably
> would add the "<rdf:RDF>" part). 

Do you find it strange that RSS 1.0 has the same properties? If you
parse the triples into a triplestore and reserialize it, you'll
typically get something which is no longer RSS 1.0. (Go ahead, try and
take RSS 1.0 ntriples and pass them to an RSS reader. I dare you. :)  

> And is serialising DOAP in N-Triples
> form of RDF an invalid operation (in the sense that it won't be valid
> DOAP) ?

Yep. It's still valid RDF. It's not valid DOAP.

Regards,
-- 
Christopher Schmidt
Web Developer