See also: IRC log
<scribe> ACTION: EricP to publish xml format CR pending OK of the results schema [CONTINUES] [recorded in http://www.w3.org/2007/09/04-dawg-minutes.html#action01]
<scribe> ACTION: LeeF to solicit endpoints for HTTP and SOAP implementations of the SPARQL protocol [DONE] [recorded in http://www.w3.org/2007/09/04-dawg-minutes.html#action02]
<scribe> ACTION: Orri and Ivan to try the results XML schema and report back the results [DROPPED] [recorded in http://www.w3.org/2007/09/04-dawg-minutes.html#action03]
<scribe> ACTION: Eric+Andy to identify and collect mf:requires URIs and put in documentation [CONTINUES] [recorded in http://www.w3.org/2007/09/04-dawg-minutes.html#action04]
<scribe> ACTION: ericP to answer "EBV of invalid numeric literals" email from Arjohn [CONTINUES] [recorded in http://www.w3.org/2007/09/04-dawg-minutes.html#action05]
<scribe> ACTION: ericP to incorporate text to fix EBV of illformed numeric literals bug in spec [CONTINUES] [recorded in http://www.w3.org/2007/09/04-dawg-minutes.html#action06]
<scribe> ACTION: ericP to poke IETF folks about registering SPARQL media types (esp. application/sparql-query) [CONTINUES] [recorded in http://www.w3.org/2007/09/04-dawg-minutes.html#action07]
<scribe> ACTION: ericP to write explanatory text saying that, like xsd:inteter and xsd:dateTime, the relative order of simple literals and xsd:strings is not defined here [CONTINUES] [recorded in http://www.w3.org/2007/09/04-dawg-minutes.html#action08]
<AndyS> I looked at the syntax tests and indeed begee's comments all looked right.
minutes from last week: http://lists.w3.org/Archives/Public/public-rdf-dawg/2007JulSep/att-0115/28-dawg-minutes.html
Minutes from last week: http://lists.w3.org/Archives/Public/public-rdf-dawg/2007JulSep/att-0115/28-dawg-minutes.html
Proposed: approve last week's minutes
ericP: looks good
resolved to approve last week's minutes
next meeting: 11 Sep, same time and place
next meeting's goal: look at draft implementation report, discuss what needs to be done to it to show interoperable implementations
LeeF: Status of IETF poking, given goal of PR by end of month
ericP: will look into how to manage schedule
<iv_an_ru> (Or my knowledge of schema is too poor to get a subtle bug).
LeeF: Do we need versioning for new test suite version?
ericP: we know most of the people already, and have to lean on the others anyway (for incomplete results), so don't think we need to version
<iv_an_ru> IMHO it's enough to place it in different path on W3C site.
AndyS: does EARL have a date of the test run?
<iv_an_ru> OK, I can accomulate different versions in the test service. It let choose the suite to run.
<group finds dc:date on each test run>
AndyS: Could use dc:date information for versioning
<iv_an_ru> dc:dateTime could be enough.
-> negative syntax tests from Benjamin
1) group graph pattern was not closed
3) same problem
4) http://www.w3.org/2001/sw/DataAccess/tests/data-r2/syntax-sparql4/syn-bad-UNION-breaks-BGP.rq
<AndyS> Also wrong: syntax-pat-01.rq
<AndyS> http://www.w3.org/2001/sw/DataAccess/tests/data-r2/syntax-sparql1/syntax-pat-01.rq
AndyS: Above test has an invalid URI
<AndyS> Has illegl URI : http:/example, not http://example
PROPOSED: Approve the 4 syntax tests from Benjamin, as fixed by Andy, modulo comments added by Eric
ericP seconds
resolved
PROPOSED: http://www.w3.org/2001/sw/DataAccess/tests/data-r2/syntax-sparql1/syntax-pat-0{1-4}.rq be reapproved with fixed URIs in PREFIX clause
AndyS seconds
resolved
<scribe> ACTION: LeeF to update approval information on all tests approved and re-approved today [recorded in http://www.w3.org/2007/09/04-dawg-minutes.html#action09]
ARQ, Algae, Glitter all pass
<iv_an_ru> Yes, I'm online.
<chimezie> I haven't run those tests yet
<iv_an_ru> We do not pass all
<iv_an_ru> We will neve pass all, due to intentional deviations from SPARQL spec.
<AndyS> This is just for the 3 new sameTerm tests?
PROPOSED: Approve the 3 sameTerm tests added by Eric (ref: http://lists.w3.org/Archives/Public/public-rdf-dawg/2007JulSep/0110.html )
<iv_an_ru> I've not yet implemented sameterm.
<iv_an_ru> But I will, so let's approve it.
ericP seconds
resolved
-> invalid turtle data et al. from Arjohn
-> AndyS claims to have fixed them
Eyeballing http://www.w3.org/2001/sw/DataAccess/tests/data-r2/dataset/dataset-11.ttl
Used to have whitespace separated brackets ( [ ] )
<iv_an_ru> who cares, frankly speaking....
http://www.w3.org/2001/sw/DataAccess/tests/data-r2/i18n/normalization-01.ttl
PROPOSED: Reapprove all tests that use the data files mentioned in http://lists.w3.org/Archives/Public/public-rdf-dawg/2007JulSep/0114.html
chimezie seconds
resolved
-> Andy's latest on the timezone issue
AndyS: To compare two dates,
compare their initial points (as per F&O), implicitly turn
them into dateTime
... dateTimes are compared when their is and isn't a
timezone
... using a worst case comparison
Andy's proposal is to fix the results for open-world/date-01
both tests are mf:requires <http://www.w3.org/2001/XMLSchema#date>
they also mf:requires mf:xsd-date-eq-xsd-dateTime
<AndyS> -> http://www.w3.org/TR/xmlschema-2/#dateTime-order
<iv_an_ru> SPARQL-to-SQL front ends don't like non-deterministic behaviour so they will elect different timezones :(
<AndyS> http://www.w3.org/TR/xpath-functions/#func-dateTime-equal
<iv_an_ru> No datetimes w/o timezones in most of SQL servers.
<AndyS> Their example is "2000-01-01T12:00:00 <> 1999-12-31T23:00:00Z" is "indeterminate" (their word)
<iv_an_ru> Maybe we should be able to specify via SPARQL query protocol: calculate this assuming the specified timezone for all timezoneless constants in query text.
ericP: could add editorial text that there might be a local timezone in effect that affects interpretation of dates
AndyS: might not need it since we rely on F&O text and subsidiaries
<scribe> ACTION: LeeF to note mf:requires of extension points when emailing out update to test suite [recorded in http://www.w3.org/2007/09/04-dawg-minutes.html#action10]
<discussion - probably should leave date-1 as an unapproved test, as it's an extension test anyway>
(meeting extended for 30 minutes)
-> AndyS on <=
ericP: I think that's consistent with XPath/XQuery
AndyS: They define their <= slightly differently, so it's not the same
issue is "1"^^my:int <= "1"^^my:int
which right now is a type error
<AndyS> so the issues arises because the defn of <= says "numeric-equals or numeric-less-than" and not ...
<AndyS> "any equal or numeric-less-than" i.e. we jumped to numerci-equals and not use our own equals (=)
A <= B
RDF Terms
<AndyS> logical-or is our own || operator
A = B || A < B
<iv_an_ru> Can we violate transitivity of '<' if we use a mix of types?
<Souri> error || true (or in other order) => what?
<AndyS> It is bizaar is A = B and then A <= B is an error
<iv_an_ru> I'd say true
<AndyS> Souri - see http://www.w3.org/2001/sw/DataAccess/rq23/rq25.html#evaluation
"""
Change the test from:
FILTER ( ?v1 <= ?v2 || ?v1 > ?v2 )
to:
FILTER ( ?v1 = ?v2 || ?v1 < ?v2 || ?v1 > ?v2 )
"""
<AndyS> That is acceptable.
<iv_an_ru> I guess taht we're no in the state when any change is critical. Finally, we do not add run-time errors to existing code if we replace <= with < || = .
<iv_an_ru> If some existing code return results it has high chances to continue returning results.
<chimezie> we would need to do the same for the other operators as well if we changed the spec, right? i.e., add entries for <= as well as >=
Orri: Avoid going to LC
Souri: not too strong feelings, avoid LC is a flash point, go with group consensus
AndyS: Change the test to be explicit about what it's testing for
ericP: change the test
chimezie: change the test
<iv_an_ru> No strong opinion. Let's go to LC and start SPARQL 2.0 with SPARUL and other extensions :)
<iv_an_ru> but change the test first.
PROPOSED: Change open-world/open-cmp-02.rq from "FILTER ( ?v1 <= ?v2 || ?v1 > ?v2 )" to "FILTER ( ?v1 = ?v2 || ?v1 < ?v2 || ?v1 > ?v2 )"
ericP seconds
resolved
Souri: suggests putting ?v1 = ?v2 in the middle to emphasize that order doesn't matter
PROPOSED: To update the test suite with all changes from today's telecon, and re-advertise to the world
<ericP> second
<iv_an_ru> agree
resolved
<AndyS> Test open-cmp-02; Changed to "FILTER ( ?v1 < ?v2 || ?v1 = ?v2 || ?v1 > ?v2 )" and test results corrected
ericP: i have data from test results and queries that look at them
<scribe> ACTION: ericP to have first cut of shiny implementation report by 11 Sep [recorded in http://www.w3.org/2007/09/04-dawg-minutes.html#action11]
<scribe> ACTION: ericP to try to produce a diff from CR to current editor's draft [recorded in http://www.w3.org/2007/09/04-dawg-minutes.html#action12]
<iv_an_ru> (I'm trying to install the service on publicly available host ATM. It's ugly but at least it demonstrates our own mismatches with the spec.)
<iv_an_ru> In addition we have one i18n issue.
<iv_an_ru> ATM I'm testing it with sparql.org and our own hosts.
LeeF: Can we publish a list of tests that it doesn't handle right?
Orri: Yes
<scribe> ACTION: Orri and Ivan to publish URL to test service along with a list of tests that aren't handled correctly yet [recorded in http://www.w3.org/2007/09/04-dawg-minutes.html#action13]
<iv_an_ru> yes, it distinguish between failed and not executed tests.
<chimezie> I have to run, unfortunately...
Lee and Eric and anyone else interested meet tomorrow (5-Sep) at 11:00AM EDT to discuss XSD problems and get to the bottom of it
adjourned