- From: Sandro Hawke <sandro@w3.org>
- Date: Tue, 20 Nov 2012 09:12:10 -0500
- To: Mike Dean <mdean@bbn.com>
- CC: Benjamin Grosof <BenjaminG@vulcan.com>, "Michael Kifer (kifer@cs.sunysb.edu)" <kifer@cs.sunysb.edu>, Chimezie Ogbuji <chimezie@gmail.com>, Paul Fodor <pfodor@cs.stonybrook.edu>, public-rdf-dawg-comments@w3.org
[Replying on the archived list, for WG attention.]
On 11/12/2012 05:53 PM, Mike Dean wrote:
> rif02 is a similar query, but uses RIF in RDF [2] instead of XML for the RIF rules [3]. We can load it as triples, but don't do anything with them. We could probably create a couple rules to handle this specific case, but that seems like cheating.
>
> Several questions:
>
> 1) Does successfully executing 1 of 2 tests meet the requirement?
>
> 2) Is it appropriate for rif02 to use RIF in RDF at CR, since that's still a Working Draft?
Thanks for catching this, Mike. As you suggest, RIF-in-RDF is a RIF
Working Group Note (not a Rec-Track document), so it's not appropriate
to require implementing it for implementing the SPARQL RIF Core
Entailment Regime.
I think the text in the spec could probably be clearer. It says: "the
URI for a RIF rule set (which may be encoded in RIF-XML or RIF-in-RDF)",
without clarifying that implementation do not have to handle both.
So, I think we'll need to make that test case optional.
At ISWC I talked to you about the possibility of putting together a few
more tests. Any chance you've had the chance to do that?
-- Sandro
Received on Tuesday, 20 November 2012 14:12:31 UTC