Re: Multiple GRDDL results in a single transform??? GRDDL and Named Graphs

Harry Halpin wrote:
> Booth, David (HP Software - Boston) wrote:
>> There are two separable issues here:
>>
>> 1. Whether test #grddlonrdf conforms to the GRDDL spec as written.  AFAICT it does *not* conform to the GRDDL spec as written, so in my view the working group has a responsibility in an erratum to either fix it or delete it.
>>   
> Could you clarify exactly why the test violates the spec as written? I
> agree, if it violates the spec then this should be noted in erratum..

I think I have partially done this on David's behalf.

I have just shown, I believe, that the comments associated with the two 
tests are incorrect.

David has previously argued that the use of #rule_rdfxbase in the two 
tests is incorrect, since the presence of the grddl:transformation 
attribute on the rdf:RDF element is not permitted by the RDF Syntax 
specification, and so the input documents are not RDF/XML despite 
initial appearance. I believe that argument to be correct but in an 
overly pedantic way (sorry David).

Given that we could, and in my view should, fix the comments in two 
(closely related) errata, we could also consider David's argument, and I 
would hope that making the tests informative may be an easy way to 
acknowledge David's argument while keeping everyone else on board, (at 
least me!)

A very minimal entry in the erratum document would be the two comments 
are misleading, possibly suggesting a textual correction of deleting the 
comments. This would not address David's point though.

Jeremy

Received on Friday, 25 January 2008 12:58:51 UTC