Update to RDFa Test Suite concerning exclusive canonical XML literals

Hi all,

Per my action item from last week's RDFa telecon[1], the RDFa Test Suite
has been updated to utilize Exclusive Canonical XML Literals when
performing the SPARQL queries. This means that the test suite is more
restrictive now and expects the SPARQL engine to properly canonicalize
RDF/XML graphs before running the SPARQL query on the graph.

This discussion and decision was triggered by Philip Taylors comment
noting that RDFa processors should always produce exclusive canonical
XML Literals. Shane wrote up some errata noting that this will be a
requirement in the future. Ivan then mentioned, in last week's telecon,
that we shouldn't require this because RDF/XML is supposed to be
canonicalized when it is fed into a graph which is then given to a
SPARQL engine. Requiring c14n would be an unnecessary burden on RDFa
processors that output RDF/XML.

Shane is currently drafting text to strongly urge RDFa processor writers
to produce exclusive canonical XML Literals (perhaps making this a
SHOULD in the future?), but not require it (not a MUST). The RDFa Test
Suite has been updated to use exclusive canonical XML Literals in the
SPARQL queries for tests that generate XML Literals:

http://github.com/msporny/rdfa-test-suite/commit/4ecfc3b287ace346ebea1c0acc0f89a782f30512

What this means for RDFa Processor developers:

If your processor does not generate exclusive canonical XML Literals
then some of the tests will start FAILing your processor. You don't have
to change anything now. We believe that there is a bug in the SPARQL
endpoints that are causing the tests to fail. You may want to change
your processor to always generate exclusive canonical XML Literals, but
you don't have to at this point in time.

What is going to be done about the failing tests?

The tests are failing because the SPARQL endpoints are not
de-serializing the RDF/XML given to them into exclusive canonical XML
Literals in the graphs that are being produced for the SPARQL query
engine. I have an action to contact Andy Seaborne and see if he knows
why the canonicalization isn't happening when the SPARQL query engine is
building the RDF graph from the RDF/XML that the RDFa Processors are
generating.

-- manu

[1]http://www.w3.org/2009/10/15-rdfa-minutes#ActionSummary

-- 
Manu Sporny (skype: msporny, twitter: manusporny)
President/CEO - Digital Bazaar, Inc.
blog: Establishing an Open Digital Media Commerce Standard
http://blog.digitalbazaar.com/2009/09/28/a-digital-content-commerce-standard/

Received on Wednesday, 21 October 2009 02:50:47 UTC