W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > October to December 2004

let's approve 8 tests: triple-pattern-001 thru example-2-4a

From: Dan Connolly <connolly@w3.org>
Date: Sun, 28 Nov 2004 18:54:37 -0600
Message-Id: <3E141209-41A1-11D9-B5C9-000D9338C596@w3.org>
To: public-rdf-dawg@w3.org

While working on SOURCE definitions...

   Formalizing SPARQL rules
   Noodling on using MathML for the definitions in the SPARQL spec, and 
deriving N3 from it.
   http://www.w3.org/2001/sw/DataAccess/mathml-rules.xml

... I ended up with a sort of implementation of a tiny bit of SPARQL in 
N3

I got it to pass the 1st test...

   http://www.w3.org/2001/sw/DataAccess/tests/#dawg-triple-pattern-001

i.e. from these data
   http://www.w3.org/2001/sw/DataAccess/tests/data/simple/data-01.n3
and this query
   http://www.w3.org/2001/sw/DataAccess/tests/data/simple/dawg-tp-01.rq
my implementation agrees that these are solutions
   http://www.w3.org/2001/sw/DataAccess/tests/data/simple/result-tp-01.n3

My implementation can't recover the information that's lost when some 
variables
are not SELECTed, but I eyeballed these and I propose we adopt them:

http://www.w3.org/2001/sw/DataAccess/tests/#dawg-triple-pattern-002
http://www.w3.org/2001/sw/DataAccess/tests/#dawg-triple-pattern-003
http://www.w3.org/2001/sw/DataAccess/tests/#dawg-triple-pattern-004
http://www.w3.org/2001/sw/DataAccess/tests/#sparql-query-example-2-1a
http://www.w3.org/2001/sw/DataAccess/tests/#sparql-query-example-2-2a
http://www.w3.org/2001/sw/DataAccess/tests/#sparql-query-example-2-3a
http://www.w3.org/2001/sw/DataAccess/tests/#sparql-query-example-2-4a

I'm not prepared to endorse the datatype tests; I don't like the
special syntax for <. I probably won't object if somebody else proposes 
it,
but I'd abstain.

There are still various issues with the test documentation:

The result format, in particular the :value property uses,
rather than mentioning, terms, but I can live with that, I suppose,
but it's similar to the problem with rdf:subject/predicate/object
that makes them useless for many things that people want to
use them for.

There's no specification for the format the test data is written in, 
i.e. turtle.
Let's at least add a link to 
http://www.ilrt.bris.ac.uk/discovery/2004/01/turtle/
for now; eventually I think a copy is in order (ala the way n-triples is
specified in the RDF Core test docs).

And I don't mean to undo our agreement to provide all the materials in 
RDF/XML
as well as turtle.
(cf 2Nov minutes
http://lists.w3.org/Archives/Public/public-rdf-dawg/2004OctDec/0210.html
and the subsequent minutes which show SteveH's continuing action).

I propose we adopt these 8 tests. Can I get a second?
DaveB, you have run them in the past; wanna double-check them
in their present condition? Jos, are you happy with these 8 tests?
Has anybody else either eyballed them or got code that passes them?

Anybody see any reason why we should *not* adopt them? Does anybody
have a problem with the syntax used in these first few tests?

I welcome ammendments to this proposal that regard only these 8 tests.
Anybody that wants to approve other tests is more than welcome to
make a separate proposal.

-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/
Received on Monday, 29 November 2004 00:54:39 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:15:21 GMT