W3C home > Mailing lists > Public > public-rif-wg@w3.org > December 2008

Re: test - metadata format

From: Stella Mitchell <cleo@us.ibm.com>
Date: Mon, 1 Dec 2008 10:56:14 -0500
To: Sandro Hawke <sandro@w3.org>
Cc: public-rif-wg@w3.org
Message-ID: <OF75AEC7F4.D141282F-ON85257512.00562051-85257512.00578BF6@us.ibm.com>
> > > >Should we be providing a traditional XML form of the Manifest?
> > > 
> > > The Manifest contains meta data about the test and uses meta data 
> > > vocabularies such as Dublin Core (DC). Do you mean with traditional 
> > > XML form that we should invent our own XML vocabulary or do you mean
> > > XML/RDF representation of DC?
> 
> I'm not sure about the metadata.  I'm just thinking people should be
> able to implement the test harness without RDF/XML.  Perhaps I'm being
> over-cautious there.  I remember asking the WG for guidance on this once
> before, and I think the WG said RDF/XML was okay, but ...  I'm having
> second thoughts, as I think about possible implementors.
> 
> > Providing the manifest in "plain" XML sounds
> > good. Did you mean instead of or in addition to 
> > the RDF/XML? Instead of seems reasonable? The group
> > never did discuss design constraints or pros/cons 
> > for different formats of the metadata. Early on 
> > in group discussions, some people thought the
> > metadata should be represented as RIF
> > annotations, and you thought it should be as close
> > to the RDF/OWL1 test ontologies and formats as possible.
> > Later, the idea of aligning with OWL2 test suite came 
> > up - their metadata is not based on RDF/OWL1 format,
> > because they want to have it be OWL-DL compatible.
> 
> Yeah, I'm having a hard time knowing the right option here.  Too many
> possibilities and considerations. 
> 
> Maybe both:
> 
>     bld-tests.rdf
>             --- all the metadata and test data, in rdf/xml
>     bld-tests.xml
>             --- all the test data, some metadata, in simpler xml
>

You gave a reason for providing it in XML, to make it simpler for a 
broader group of test harness implementers. What would be the reason to 
also provide it in RDF/XML? How about the file representing all the tests 
having all the test data and all the metadata, and then for each 
individual test, all the metadata only?

Stella
 
> > > -Adrian
> > > 
> 
>     -- Sandro
> 
> > > >      -- Sandro
> > Stella 
Received on Monday, 1 December 2008 15:56:56 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:33:59 GMT