W3C home > Mailing lists > Public > public-grddl-wg@w3.org > February 2007

Re: #grddlonrdf test-case: approve?

From: Chimezie Ogbuji <ogbujic@bio.ri.ccf.org>
Date: Sat, 10 Feb 2007 22:07:36 -0500 (EST)
To: Harry Halpin <hhalpin@ibiblio.org>
cc: public-grddl-wg <public-grddl-wg@w3.org>, dave@dajobe.org, jjc@hpl.hp.com
Message-ID: <Pine.GSO.4.60.0702102146180.27962@joplin.bio.ri.ccf.org>

On Sat, 10 Feb 2007, Harry Halpin wrote:
> Now, this is non-trivial, as we could propose to accept only 1 or any
> combination of two of the above test-cases. If we do accept 1), for
> example, then GRDDL results are non-additive. However, I do think it
> makes sense, since I guess quite a few implementations are going to do 1)

The way the GRDDL specification is (currently) written you can't mandate 
any of the options - it doesn't *require* the merging of all possible 
GRDDL results or any specific combination, it only defines each of the 
three as 'valid' GRDDL results.  This makes testing for compliance a bit 
more complicated (as we've seen with the multiple output test scenarios as 
of late) but leaves implementors with the ability to follow their 
own policies.

While on the subject, it seems that there is some ambiguity in whether or 
not profile GRDDL results are GRDDL results of the source documents that 
reference them

It led to some recent confusion in #swig:

<dajobe> see I get 136 triples for that, 'cos I import all of the profile document's triples into the main card.html doc.
<chimezie> the profile document's triples should be a GRDDL result of that URI, no?
* chimezie recalls some @@ indication about that in the spec
<chimezie> huh.. when did that change?
<dajobe> I'm not clear, I might have coded it wrong.
<chimezie> i dont think you did actually..
<chimezie> [[[
<chimezie> if an information resource ?D has an XHTML representation whose profile attribute refers to ?PROFILE, then any GRDDL result of ?PROFILE is a GRDDL result of ?D
<chimezie> ]]]
<chimezie> followed by:
<chimezie> @@no, profile results don't get merged in with ?D. and @@TODO: be more clear about multiple profiles.
... skip ...
<dajobe> so profile results are just scraped for ?x grddl:profileTransformation ?y triples, others are dropped
<chimezie> im just wondering what is the arch.. motivation for not including them as a result...
<chimezie> it seems to be the same for ns docs
<dajobe> they are mostly triples about different resources

Chimezie Ogbuji
Lead Systems Analyst
Thoracic and Cardiovascular Surgery
Cleveland Clinic Foundation
9500 Euclid Avenue/ W26
Cleveland, Ohio 44195
Office: (216)444-8593
Received on Sunday, 11 February 2007 03:22:47 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:39:10 UTC