See also: IRC log
Meeting start: 14:37 Z
Dec 12 minutes and Dec 19 minutes both accepted as true records (seconded by Andy in both cases).
Lee talked with Pat Hayes about the status of the entailment text, reporting that Pat expects to have second draft text for the WG to look at by next meeting (Jan 9th).
Simon posted the results of his investigation into the XML reporting format from xUnit to the comments list. (Next agendum includes summary.)
Other actions continue, with Lee taking over Kendall's outstanding actions.
Eric described his work on yacker
, a tool to match patterns in a query's abstract syntax tree. By making a list of SPARQL's language features (facets) and associating each with the grammar productions used to invoke it, test cases could be automatically annotated with the language features they cover.
Jeen: Thought that features were associated implicitly to tests by the way we group them. However if doing it explicitly is useful, see no harm in explicit association.
<ericP> http://www.w3.org/2001/sw/DataAccess/tests/facets.ttl (SPARQL's language features in RDF)
<ericP> http://www.w3.org/2001/sw/DataAccess/tests/facets (SPARQL's language features styled into HTML
<ericP> http://www.w3.org/2001/sw/DataAccess/tests/facets#WhereClause-noWhere (a particular facet for the following example)
:WhereClause-noWhere a test:DAWGfacet;
test:order 17;
rdfs:comment "WHERE^H^H^H { ... }";
test:production "WhereClause";
test:productionURI sparql:rWhereClause .
:SelectQuery-star a test:DAWGfacet;
test:order 9;
rdfs:comment "SELECT *";
test:production "SelectQuery";
test:productionURI sparql:rSelectQuery;
test:yaccProduction "_O_QVar_E_Plus_Or__QTIMES_E__C";
test:rule "TIMES" .
test#PrefixDecl
test#SelectQuery-Var1
test#GraphPattern-spo
test#WhereClause-noWhere
test#OrderCondition-rVar
...
!?s ?p []
!?x+?y
!?x-?y
<ericP> http://tinyurl.com/ybxr85 (yacker: SPARQL validation results)
ericP: Would like to annotate existing tests before migrating to new tests
AndyS: If tests have a URI, then the detail about language features associated with tests can be put in a separate file -- they're really only of interest to the WG rather than external testers
EricP: Currently facets have URIs; tests don't yet. Inclined to drop the facets into the existing tests, and give tests URIs during the changeover
<scribe> ACTION: EricP to run the yacker tool over and annotate the existing tests [recorded in http://www.w3.org/2007/01/02-dawg-minutes.html#action01]
Jeen: Can make a start at some of the less controversial tests to get the process shaken down
EricP: Talked about using EARL as the reporting format; is Jeen interested in doing that?
Jeen: slightly reluctant since EARL is new and different from the current tests. What's the benefit?
<LeeF> ACTION: Jeen to get started setting up new test directory and picking some initial tests for the working group to approve on 9 Jan [recorded in http://www.w3.org/2007/01/02-dawg-minutes.html#action02]
Simon gave a summary of his posting to the comments list about xUnit as a possible alternative to EARL. When JUnit runs it outputs an XML summary of the tests that failed and passed -- other tools can consume this XML (e.g. automated build tools). There are a couple other communities in python and maybe C++ that are using this, but it hasn't become a (defacto) standard yet. If we wanted to take advantage of JUnit-based tools, we could query the EARL test reports with SPARQL and XSLT that to JUnit's XML format. JUnit doesn't provide an official schema.
<ericP> LeeF: seems we wouldn't gain anything over EARL, so lets stick close to our roots
<ericP> +1
LeeF: Defer discussing the algebra until Fred is available
Andy has the text to be added to the spec in order to enact the decision about the bnode/cardinality issue.
LeeF: In a previous thread, it was agreed that we could live without commas in the SELECT clause.
EricP: Does LeeF see a path towards mandatory commas?
LeeF: Once SPARQL 1 is released, unlikely.
<LeeF> PROPOSED: To close http://www.w3.org/2001/sw/DataAccess/issues#punctuationSyntax issue, with no change (no commas in the SELECT list or ORDER BY list)
Seconded by Jeen with no objections or abstentions
<LeeF> ACTION: LeeF to update the issues list marking http://www.w3.org/2001/sw/DataAccess/issues#punctuationSyntax closed [recorded in http://www.w3.org/2007/01/02-dawg-minutes.html#action03]
Next meeting: Jan 9 (No scribe nominated yet)
Meeting adjourned at 15:30.