See also: IRC log
<kendall> Rachel Yager, Financial Services Technology Consortium
<kendall> new group member?
<ericP> Time: 2005-09-06T14:30Z
-> http://lists.w3.org/Archives/Public/public-rdf-dawg/2005JulSep/att-0320/Aug30.html minutes 30 Aug
RESOLVED to approve
<kendall> Minutes from last week's meeting approved
<ericP> Regrets for next week: JosD
<scribe> ACTION: DanC to find pointers to consideration of splitting CONSTRUCT / SELECT into separate interfaces [recorded in http://www.w3.org/2005/09/06-dawg-minutes.html#action01]
note WG is not persuaded by MarkB's arguments to reconsider protocol design
<scribe> ACTION: KC to respond to inconsistent parameter names commentor [recorded in http://www.w3.org/2005/09/06-dawg-minutes.html#action02]
http://www.w3.org/2001/sw/DataAccess/proto-wd/ 1.67
<LeeF> http://www.w3.org/2001/sw/DataAccess/proto-wd/#query-bindings-http
<LeeF> I think.
<scribe> ACTION: [DONE] KendallC to add editorial note to protocol spec showing that our WSDL is not kosher and it depends on their final decision. [recorded in http://www.w3.org/2005/09/06-dawg-minutes.html#action03]
<scribe> ACTION: [DONE] KendallC, to relax the query-result type to allow "equivalent serialization" [recorded in http://www.w3.org/2005/09/06-dawg-minutes.html#action04]
in 2.1.3 query Out Message
c. CONSTRUCT with simple RDF dataset and HTTP content negotation
<scribe> ACTION: [DONE] KC to make conneg explicit in c. CONSTRUCT with simple RDF dataset and take accept: out elsewhere [recorded in http://www.w3.org/2005/09/06-dawg-minutes.html#action05]
KC: example numbering may change, and the example content will be generated by Lee's code, but otherwise, the examples are done
"finish description of SOAP bindings"
<DaveB> there's about 4 '@@' labels in proto-wd as of now
<kendall> I believe all of the @@ markers are for section numbers.
PROPOSED: to publish http://www.w3.org/2001/sw/DataAccess/proto-wd/ as Last Call WD, v 1.67 plus example section numbers, plus "finish description of SOAP bindings" reviewed by @@
<kendall> 1.68 has the proper compliance language, FWIW.
PROPOSED: to publish http://www.w3.org/2001/sw/DataAccess/proto-wd/
as Last Call WD, v 1.67 plus example section numbers, plus
fixing @@ section refs, plus "finish description of SOAP
bindings" reviewed by @@
... to publish http://www.w3.org/2001/sw/DataAccess/proto-wd/
as Last Call WD, v 1.68 plus example section numbers, plus
fixing @@ section refs, plus "finish description of SOAP
bindings" reviewed by @@
<kendall> A compliant SPARQL Protocol service must support the SparqlQuery interface; if a SPARQL Protocol service supports either HTTP or SOAP bindings, it must support the bindings as described in sparql-protocol-query.wsdl. A SPARQL Protocol service may support other interfaces.
<DaveB> and k. elides the body of the query with '...'
PROPOSED: to publish http://www.w3.org/2001/sw/DataAccess/proto-wd/ as Last Call WD, v 1.68 plus example section numbers, plus fixing @@ section refs, plus "finish description of SOAP bindings" reviewed by @@, plus change "i. This SPARQL query" to "i. This erroneous SPARQL query..."
<kendall> I added "syntactically invalid"
<afs> The current protocol draft + agreed mods as per Dan's proposal + clearer bad syntax is good to me.
so we'll regen the examples after last call.
PROPOSED: to publish http://www.w3.org/2001/sw/DataAccess/proto-wd/ as Last Call WD, v 1.68 plus example section numbers, plus fixing @@ section refs, plus "finish description of SOAP bindings", plus change "i. This SPARQL query" to "i. This erroneous SPARQL query..."
<kendall> "syntactically invalid" == "erroneous"?
sure.
<kendall> cool
so RESOLVED.
<scribe> ACTION: EricP to publish proto-wd as above [recorded in http://www.w3.org/2005/09/06-dawg-minutes.html#action06]
critical path: EP, KC
(yes, those actions under item 12 are continued)
<scribe> ACTION: ericP to send [OK?] message to Bjoern. re BASE IRI resolution comment [recorded in http://www.w3.org/2005/09/06-dawg-minutes.html#action07]
(DanC will explain more after the meeting)
lang(?x) = "en" # or not
lang-match("en-us", "en") # true
lang-match(?, "en") # true if ?x is bound to "en-us"
<DaveB> thinking: so for some rdf literal ?x, lang-match(lang(?x), "en") is true if for example, ?x is bound to "EN-US"
<ericP> propose:
<ericP> lang-match(lang(?x), "en") matches if ?x@en or ?x@en-us
<ericP> lang-match(lang(?x), "en-us") matches if ?x@en or ?x@en-us
<ericP> .
<afs> Minor: langMatch (no "-") would be more inline with the grammar so far. Could have a keyword with "-" no probs there.
<scribe> ACTION: ericp to draft lang-match design, summarizing and citing RFC3066 [recorded in http://www.w3.org/2005/09/06-dawg-minutes.html#action08]
<ericP> DanC: bonus points for test cases
yes, it would help to have "known datatypes" in the test manifest (as in WebOnt/OWL)
<afs> Round table?
EricP pref 0
DanC pref 1
Jos pref 1
Andy pref 0... er... 0 or 1 depending on server knowledge
<afs> "=" can only return true if it positively knows the fact to be true (and it can be odd that !(XXI=21) \=> XXI!=21)
<DaveB> I say, if we say the list of datatypes supported is (... a set ..) and that set is the set of datatypes in the query spec, the answer must be 0
<afs> Can URIs denote numbers?
<ericP> experiments with flags in result sets
<ericP> modifying truth table for OR
(maybe I should have reopened issues#disjunction instead of or in addition to #valueTesting)
EricP: currently, (typeerror || X) give X. commentor asks for (typerror || true) gives true, (typerror || false) gives typeerror [did I get that right, ericp?]
<afs> issue #disjunction is about graph pattern disjunction but here we are in F&O space. I don't know if these can be equated.
ok, so #disjunction is separate.
<afs> That's as I understand it. It would be nice if filters and graph patterns were more strongly related.
<afs> But it may be that it just isn't possible to use F&O and equate to graph patterns in all cases
(chair is at that point of choosing between putting a question or actioning somebody to make the WG smarter)
<ericP> extendedType-eq-pass and extendedType-ne-fail
<DaveB> ?? !("a" 1) || true
<DaveB> ?? !("a" + 1) || true
<ericP> !(typeError || FALSE)
<afs> !(||) is related to && Do these the truth tables work out for && as well?
<ericP> !("a" + 1 || FALSE)
<ericP> by LC, that returns a TRUE
<ericP> by TimBL, it returns no solution
<ericP> (i.e., the FILTER encounter a typeError and eliminates that solution)
<ericP> modifying truth table for OR
<scribe> ACTION: EricP to get SteveH's response to 0360 modifying truth table for OR [recorded in http://www.w3.org/2005/09/06-dawg-minutes.html#action09]
<afs> I *think* that's OK - would like to test before vote
<ericP> ditto
AFS: ORDER BY devolves to < ... which isn't defined for IRIs, strings...
<DaveB> (fn:compare has an optional collation argument)
<DaveB> http://www.w3.org/TR/xpath-functions/#func-compare
<kendall> cool
PROPOSED: to extend < to apply
to strings and IRIs, and to use this for sorting IRIs.
... to extend < to apply to strings and IRIs based on
XPath/XQuery F&O #func-compare, and to use this for sorting
IRIs.
<afs> Q for understanding: Why "<" on IRIs? Use case?
PROPOSED: to extend < to apply
to strings based on XPath/XQuery F&O #func-compare, and to
use this for sorting IRIs.
... to extend < and other relational ops to apply to strings
based on XPath/XQuery F&O #func-compare, and to use this
for sorting IRIs.
<DaveB> ref http://www.w3.org/TR/xpath20/#static_context
<DaveB> in http://www.w3.org/TR/xpath-functions/#collations there is the reference to the defaults for collation and the unicode code point collation
PROPOSED: to extend < and other relational ops to apply to strings based on XPath/XQuery F&O #func-compare, and to use this for sorting IRIs.
so RESOLVED.
<scribe> ACTION: DanC to ask I18N WG for review [recorded in http://www.w3.org/2005/09/06-dawg-minutes.html#action10]
<scribe> ACTION: EricP to extend < and relational ops, get review by Andy [recorded in http://www.w3.org/2005/09/06-dawg-minutes.html#action11]
<scribe> ACTION: DaveB to find query with FILTER on URIs [recorded in http://www.w3.org/2005/09/06-dawg-minutes.html#action12]
action -9
<kendall> I changed <query><sparql-query>SELECT to <query><query>SELECT, and I should have changed it to <query-request><query>SELECT
<afs> That looks fine.
PROPOSED: to change the outer <query> element name to <query-request>
so RESOLVED. EricP abstaining
http://www.w3.org/2001/sw/DataAccess/lc-status-report.html
ADJOURN.