W3C home > Mailing lists > Public > public-rdf-dawg-comments@w3.org > September 2007

Re: SPARQL test suite updated - soliciting test suite results

From: Gregory Williams <greg@evilfunhouse.com>
Date: Wed, 5 Sep 2007 20:07:19 -0400
Message-Id: <4250332B-1E64-4D83-ACD4-2573FD7D22B7@evilfunhouse.com>
To: public-rdf-dawg-comments@w3.org

In the test <http://www.w3.org/2001/sw/DataAccess/tests/data-r2/ 
distinct/manifest#distinct-9>, there is an expected result with the  
binding v = "1.3e0"^^xsd:double. The data for this test is specified  
in turtle, with triples such as:

:z3 :p1 1.3e0 .

However, the turtle spec allows for any legal lexical representation  
of an xsd:double in parsing unquoted double values[1]. Given this, it  
is possible for the :distinct-9 query to return valid results with  
any number of lexical representations for 1.3e0, causing  
discrepancies between the actual results and those listed in the test  
suite.

Does any document relating to the test suite or the SPARQL query  
results xml format discuss this issue? I may have missed something,  
but given the definition of graph equivalence[2] and literal equality 
[3], I believe this is a problem with the test data. Is a test  
harness meant to be aware of all possible lexical representations of  
these values? Any feedback on this issue would be greatly appreciated.

thank you,
gregory williams


[1] http://www.dajobe.org/2004/01/turtle/#double
[2] http://www.w3.org/TR/rdf-concepts/#section-graph-equality
[3] http://www.w3.org/TR/rdf-concepts/#section-Literal-Equality

-- 
"Viking went to Mars and asked if it had life, and Mars answered
  by replying 'Could you please rephrase the question?'"
     - Leonard David
Received on Thursday, 6 September 2007 00:07:36 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:14:51 GMT