W3C home > Mailing lists > Public > public-rdf-dawg-comments@w3.org > February 2006

Re: langMatches wording in sparql query 2005-11-23 and tests [OK?]

From: Dave Beckett <dave@dajobe.org>
Date: Fri, 17 Feb 2006 16:18:45 -0800
Message-ID: <43F667E5.80807@dajobe.org>
To: Dan Connolly <connolly@w3.org>
CC: public-rdf-dawg-comments@w3.org

Dan Connolly wrote:
> On Thu, 2006-01-05 at 23:31 -0800, Dave Beckett wrote:
> 
>>The langMatches definition:
> 
> [...]
> 
>>I think this is backwards, as an example following uses '*' in the
>>second argument position, and so do the test cases I was puzzling over -
>>q-langMatches-[1-4].rq
> 
> 
> quite. fixed in
> http://www.w3.org/2001/sw/DataAccess/rq23/#func-langMatches
> 1.643 of 2006/02/17 23:04:04
> 
> Please let us know if this comment addresses your questions to
> your satisfaction.

Yes, the change in 1.643 addresses my comment.

> 
>>I'm also unsure about the tests as q-langMatches-3 returns an answer
>>with no language but the definition above says * only matches non-empty
>>languages. q-langMatch-4 which is the negative of that, is similarly
>>affected.
> 
> 
> I hope it's OK if I leave your question on the test materials aside,
> for now. We'll be going over them in CR.

OK, to keep this tracked for the future please record this thread as a
test material comment.

> 
> 
>>The tests also rely that LANG(non-literals) returns "" although it's not
>>so explicit in the text to me as LANG() is only defined for literals,
>>including in the table "SPARQL Unary Operators".  I was returning a type
>>error taking anything not allowed as forbidden.

Dave
Received on Saturday, 18 February 2006 00:18:51 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:14:50 GMT