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

Re: Comments on dawg test cases

From: Seaborne, Andy <andy.seaborne@hp.com>
Date: Fri, 03 Nov 2006 15:56:30 +0000
Message-ID: <454B66AE.7040004@hp.com>
To: Faisal Alkhateeb <Faisal.Alkhateeb@inrialpes.fr>
CC: public-rdf-dawg-comments@w3.org

Faisal Alkhateeb wrote:
 > Subject: Comments on dawg test cases
 > Date: Mon, 16 Oct 2006 10:12:37 +0200
 > From: Faisal Alkhateeb
 > To: public-rdf-dawg-comments@w3.org

 > Hello,
 >    The result of the following query is not clear
 > http://www.w3.org/2001/sw/DataAccess/tests/#dawg-triple-pattern-005
 > and i think the result is empty.
 > Regarding the sorting of the following query, is not it in the reverse order
 > (that is descending as it is clear, since E> F> B> A).
 > http://www.w3.org/2001/sw/DataAccess/tests/#sort-1
 > And so on for the rest of sorting queries.

The recorded results are right - it looks like the HTML generator does not
respect the "index=" in the results files.  The HTML page is generated from 
the underlying test suite: http://www.w3.org/2001/sw/DataAccess/tests/data/

It would be clearer to use SRX files for ordered results - the HTML generator
doesn't seem to understand them, looking at their use elsewhere.

 > The sorting result of the following query is mixed (i.e., neither ascending
 > nor descending)
 > http://www.w3.org/2001/sw/DataAccess/tests/#sort-7


 > I want to know if it is syntactically possible to use the keyword FILTER
 > as a namespace prefix as done in the following query:
 > http://www.w3.org/2001/sw/DataAccess/tests/#keyword-filter-as-a-namespace-
 > prefix

It's legal.  Any keyword can be used as a prefix.  This happens in the
tokenizing part of the grammar because "FILTER:" is a longer match than "FILTER"

 > In the following queries, we use '(' after WHERE clause instead of '{'.
 > http://www.w3.org/2001/sw/DataAccess/tests/#dawg-unsaid-001
 > http://www.w3.org/2001/sw/DataAccess/tests/#dawg-unsaid-002

Old syntax.

The test suite is both a current and historical.  It needs to be split into 
those things that are current and approved and the rest.  These tests record 
and old syntax and old feature.  Neither in current SPARQL.

 > And finally, for the following query, the result is not empty:
 > http://www.w3.org/2001/sw/DataAccess/tests/#extendedtype-ne-fail

Open world issue.  != means not known to be different so unless the processor
knows about the types, they are not known to be different values so != does
not return true.

Received on Friday, 3 November 2006 15:56:50 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:52:08 UTC