W3C home > Mailing lists > Public > public-grddl-wg@w3.org > January 2008

Re: on test #grddlonrdf

From: Jeremy Carroll <jjc@hpl.hp.com>
Date: Fri, 25 Jan 2008 12:46:41 +0000
Message-ID: <4799DA31.1020500@hpl.hp.com>
To: "Booth, David (HP Software - Boston)" <dbooth@hp.com>
CC: "public-grddl-wg@w3.org" <public-grddl-wg@w3.org>

Booth, David (HP Software - Boston) wrote:
> I'm confused.

Oh yes you are very right.

There is definitely an error in that the text of the comment presupposes 
that the two output files were the other way round, and I had 
unconciously corrected the links to read the other way
Note that the input is an RDF document with a GRDDL transformation, and 
that according to the rules given by the GRDDL Specification, there are 
three distinct and equally valid output graphs for this test  for this 
document. This  *output* is a graph that is merge of the graph given by 
the source document with the graph given by the result of the GRDDL 

The word output is linked to


which has the single triple, whereas the 'merge of the graph[s]' is


(Which is specified as the output of the test)

So the text is written as if the links were the other way round.

But our preference for showing the maximal GRDDL result in the test 
document, argues for the current output document link, and hence 
significant change to the wording of the comment.

Concerning the (in)applicability of #rule_rdfxbase:

My implementation passes this test, as do others, because ignoring the 
minor error in the RDF is sensible (in my judgement), but formally 
speaking you are correct to say that the spurious attribute invalidates 
#rule_rdfxbase. Possibly this test ought to be informative.

I think what would be consistent with our earlier decisions would be to 
make the test informative, on the basis that it shows plausible 
implementation behaviour in the face of (slightly) ill-formed input, 
which the spec does not cover. An implementation that refused to fire 
#rule_rdfxbase would, at a formal level, be correct.


The related test

suffers the same error but worse, in that both *output* links are to

but the text of the comment suggests the first link should be

I fear an erratum would be in order.

Received on Friday, 25 January 2008 12:47:31 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 19:52:40 UTC