See also: IRC log
<kendall> Eric's $1 cellphone is fun till it starts shocking him!
<DanC> (auditing http://www.w3.org/2001/sw/DataAccess/#who ... missing Bijan. odd.)
<Yoshio> hi
who is [IBM] ?
<SteveH> jakim, whos on the phone?
<kendall> +1
<SteveH> ericP, [IBM] might be me
last weeks minutes approved
<DanC> scribe next week: KendallC/DanC
<DanC> looking at 1.31 http://www.w3.org/2001/sw/DataAccess/proto-wd/
for the record, continued actions:
<scribe> ACTION: KendallC to work out remaining details in sparqlx.{rnc,xsd,rng}: expression syntax, order-by stuff (xmlAbstractSyntax) [recorded in http://www.w3.org/2005/04/26-dawg-irc]
<scribe> ACTION: AndyS to define a term for abstract form of SPARQL query [recorded in http://www.w3.org/2005/04/26-dawg-irc]
<scribe> ACTION: PatH to internal review of rq23 [recorded in http://www.w3.org/2005/04/26-dawg-irc]
<scribe> ACTION: SteveH to prepare test cases for publication as WG Note (no deadline/urgency) [recorded in http://www.w3.org/2005/04/26-dawg-irc]
<scribe> ACTION: DanC to follow up re optional test based on op:dateTime triple [recorded in http://www.w3.org/2005/04/26-dawg-irc]
<scribe> ACTION: EricP to pair with SteveH on making the HTML test results page [recorded in http://www.w3.org/2005/04/26-dawg-irc]
<scribe> ACTION: DaveB to to propose source test to approve [recorded in http://www.w3.org/2005/04/26-dawg-irc]
<scribe> ACTION: AndyS to add the above graph test cases (analagous to valueTesting test cases) (don't expect quick delivery) [recorded in http://www.w3.org/2005/04/26-dawg-irc]
KendallC: HTTP binding spec is fluid.
attribute names changing.
... parts not dependent on that are all done.
... went with 1 default graph and one or more named graphs
<kendall> Andy's proposal, as I understood it, was for 1 default graph and for zero or more named graphs.
<JanneS2> (sorry for joining late, the unknown ID would be me)
DanC: steve, what if you get only one background graph?
<AndyS> Minor - can there be zero default graph URIs?
Steve: need to read the proposal
<AndyS> Not sure it makes a huge difference. Might default to empty??
KendallC: I understood 0 or more BG
and 0 or more named at the end of the last Telecon
... I understood Andy's modification to be a simplification
<kendall> It works for me to have zero background graphs
<scribe> ACTION: EricP to try out the "2 list of URIs" datasets design from <http://lists.w3.org/Archives/Public/public-rdf-dawg/2005JanMar/0440.html> [CONTINUED] [recorded in http://www.w3.org/2005/04/26-dawg-irc]
KendallC: I'm confident that what we express what we want in HTTP. not sure what the exact bytes are for the WSLD HTTP binding.
DanC: you use a complex type. iirc, the HTTP GET binding only specifies simple types.
<kendall> This is what I'm talking about:
<kendall> http://www.w3.org/TR/2004/WD-wsdl20-bindings-20040803/#_operation_styles
<DaveB> seems to me it's way to early to be using wsdl2 here, when it takes 10 mins to write an http binding for sparql
<DanC> ok? ACTION KendallC: ask WSD WG for clarification about HTTP binding of WSDL2, esp lists of URIs
EricP: there are two open source WSDL 2.0 impls at apache
<DanC> ACTION: KendallC to ask WSD WG for clarification about HTTP binding of WSDL2, esp lists of URIs [recorded in http://www.w3.org/2005/04/26-dawg-irc]
<kendall> btw, it's the xs:choice between SparqlX and xs:string forms of the query that is at issue
<scribe> ACTION: KendallC to own wsdlAbstractProtocol (based on WSDL description added to protocol to editor's draft). [recorded in http://www.w3.org/2005/04/26-dawg-irc]
<DanC> (blending with next agendum)
<scribe> ACTION: KendallC to incorporate EricP's privacy text, salting to taste [recorded in http://www.w3.org/2005/04/26-dawg-irc]
<kendall> this bit:
<kendall> <xs:choice>
<kendall> <xs:element ref="sparqlx:rdf-query"></xs:element>
<kendall> <xs:element ref="sparql-query"></xs:element>
<kendall> </xs:choice>
<DanC> <output messageLabel="Out" element="st:query-result">
<jeen> ack
AndyS: would like to see the protocol
spec aligned with DaveB's result set syntax
... outer tag is called <sparql>
... <result/> is used elsewhere
KendallC: will fix before publication
<Zakim> DanC, you wanted to ask for an example of a SELECT, a CONSTRUCT, and an ASK
KendallC: what's harder is that the result is a choice between to elements
DanC: i think you need two
interfaces
... make your best guess. this is what you want the community to
review.
<Zakim> Yoshio, you wanted to ask can one ask the server what is used as the background graph?
<DanC> [[
<DanC> _:factBookService
<DanC> saddle:dataSet <...fact-book-part-1.rdf>, <...fact-book-part-2.rdf>.
<DanC> ]]
yoshio: [above Q: +] can we use sadl to ask about the BG names
danc: there is a notion of a saddle data set
<DanC> yes, http://www.w3.org/2005/03/saddle/# is 404 unfortunately
yoshio: is namespace .../# acceptable?
DanC: not aware of any prob.
EricP: http://www.w3.org/2005/03/saddle/#somethingClever is serializable as a qname
<kendall> (ok, fwiw, the misalignment with results format is fixed, now...in version 1.32)
(too bad, could have used it to get good ACTION karma)
<kendall> :>
<DanC> ACTION: Yoshio to edit saddle descriptions provided by KendallC [recorded in http://www.w3.org/2005/04/26-dawg-irc]
<kendall> ACTION: kendall to uncommment saddle stuff, email to list [recorded in http://www.w3.org/2005/04/26-dawg-irc]
<DanC> ACTION: DanC to generate SELECT, CONSTRUCT, ASK protocol traces [recorded in http://www.w3.org/2005/04/26-dawg-irc]
<Zakim> ericP, you wanted to say the critical issue to me is whether i can do prov queries on the BG graph
<kendall> oh, come on, we all know you don't *really* care about rules! :>
<DaveB> I was just going to give info, I won't be using this protocol draft without an http section with examples in it
<DanC> PROPOSED: to publish protocol WD, contingent on inclusion of HTTP examples, saddle examples
DanC: any problem with publishing SPARQL protocol on Fri?
<DanC> PROPOSED: to publish protocol WD, contingent on inclusion of HTTP examples, saddle prose
DaveB: not useful to me unless there are HTTP examples.
<kendall> yoshio: w3c's rules workshop is this week, Wed & Thur, in DC.
DanC: critical path: KendallC, YoshioF, EricP, DanC
RESOLVED
<kendall> which namespace is 404?
<kendall> well, even funnier for DanC not to really care about, which is what I meant (smart-assedly, to be sure)
<kendall> ISSUE: Normativeness of the XML schemas. Pick one?
namespace <http://www.w3.org/2005/03/saddle/#> no longer a 404
AndyS: given streaming requirents, you might want to say whether it's ordered
<DaveB> monotonicalliy increasing index integers
DaveB: will add that it's monotonically increasing
<DaveB> 2.3.3 editorial note
<DaveB> 4 also re schemas
DanC: it would be nice to publish the results wiht protocol
<DanC> PROPOSED: to publish http://www.w3.org/2001/sw/DataAccess/rf1/ 1.25 + editorial notes from DaveB
RESOLVED
<scribe> ACTION: DaveB to work with EricP to clarify valueTesting proposal [CONTINUED] [recorded in http://www.w3.org/2005/04/26-dawg-irc]
<scribe> ACTION: EricP to propose a test re "x"^^mytype != "y"^^mytpe [CONTINUED] [recorded in http://www.w3.org/2005/04/26-dawg-irc]
<DanC> ACTION: DanC to send comment on non-use of <http://www.w3.org/2001/XMLSchema-datatypes> to schema comments list [recorded in http://www.w3.org/2005/04/26-dawg-irc]
ACTION: HowardK to seek clarification on valueTesting w.r.t. differenced from XQuery. continues
scribe is losing a lot of audio. please back me up by scribing important things you notice
AndyS: EricP, is the spec alligned with your head?
<AndyS> EricP: Yes it is aligned.
EricP: yes, needs expanding, not directly contradictory
<DanC> does 1.42 of http://www.w3.org/2001/sw/DataAccess/tests/ include Jeen's tests?
<scribe> ACTION: JeenB to Write a few sort test cases. <http://lists.w3.org/Archives/Public/public-rdf-dawg/2005AprJun/0159.html> [recorded in http://www.w3.org/2005/04/26-dawg-irc]
<DanC> (hardware problems... ew... that's why I want redundancy in the ability to regen the .html)
<DanC> (andy did or did not get the same answers?)
<SteveH> DanC, agreed
DanC, DONE?
<DanC> not sure yet, ericp
AndyS: i can pass some of JeenB's tests, but not all
<scribe> CONTINUED
<DaveB> (I've made 1.26 of http://www.w3.org/2001/sw/DataAccess/rf1/ with issues)
<AndyS> AndyS: Passes for the wrong reasons (failure to check indexes)
<kendall> hostile irc mutings?! :>
<jeen> (andys, checked in fix for test 1, should work now)
<AndyS> Thanks.
<kendall> -groan-
<AndyS> Pass (with weak testing only)
AndyS: i consider the grammar in bugfix mode
<kendall> triple-quoted strings, like python? """Hi, I'm a string"""?
SteveH, is there a reason to have WHERE optional?
<DaveB> I agree with steve on """s, WHERE {} compulsory
<Zakim> ericP, you wanted to ask if redundant bNodes are smushed
<AndyS> """ has an issue about line ends :-)
<DanC> what's the line-end issue?
<kendall> (just trying to make sure I heard correctly)
<DanC> yes, like python, """Hi, I'm a string"""
<kendall> thx
lots of drop out. DaveB, please scribe.
<AndyS> DanC: If we have a bare NL in a string, can be \r\n or \n -- Linux or Windows
<DaveB> just acking that """turtle style""" is ok
<DaveB> and make these things compulsory: 1) WHERE 2 ) outer {}
<DanC> yes, an both \r\n and \n have the same impact? ah... no, the don't match the same things. I see.
<DanC> PROPOSED: that the grammar in 1.326 http://www.w3.org/2001/sw/DataAccess/rq23/ addresses punctuationSyntax
<DaveB> detail detail, there's more clashes with turtle in the wd.
<DaveB> I've not read rq23 since then
DaveB: abstain or neigh
<Janne> me abstains, too, need to reread - I need to run for tonight to do baby sitting, cu next week
DanC: a lot of people want turtle alignment
<DaveB> I have no time before next telcon
DanC: makes DaveB (close to) critical path
<DanC> EricP: I got LALR [(1)?] problems with {}
<DaveB> ericP I presently have no conflicts, but I think I ripped out some bits
()
<kendall> wow, audio is awful
<DaveB> lol
<DaveB> let's see, my lexer has 4 states for IDs, prefixes and """-literals x2
<DanC> (so Andy, a hunk of java linked from grammar section would be quite resopnsive to my requests.)
<DanC> (linked fromt the grammar section and published as part of the WD)
<DaveB> I was saying I use approx. first 60 tokens in yacc, then do my own thing in lexer. turtle stuff I threw away and took from my turtle parser/lexer
<DanC> http://lists.w3.org/Archives/Public/public-rdf-dawg/2005AprJun/0154.html
ideally, i would like a spec that allowed an implementation without being very clever
<SteveH> +1
ADJOURNED
<Yoshio> oh
<SteveH> bye
<DaveB> ericP: i've done hopefully all the rf1 changes that I'm going to, over to you
roger that
<Yoshio> DaveB? what surprised me in your message was that "." can be understood as a part of a variable name, which grammar is it specified in?
<Yoshio> (^_^;)
<Yoshio> DaveB? what surprised me in your message was that "." can be understood as a part of a variable name, which grammar is it specified in? (in case you missed what I typed in just before you leave)
<DaveB> what grammar? it's in xml qnames I think
<DaveB> but n3 doesn't like it
<DaveB> and turtle currently doesn't allow it
<DaveB> we're moving to allow . in both sparql and turtle
<DaveB> and use the same xml qnames definition
<Yoshio> ah, so not yet in the documents
<AndyS> Any idea of when might there be an updated Turtle spec?
<Yoshio> hmm, I support having "." as in N3, but not like haveing it as a part of a variable name...
<DaveB> maybe end of next week but www2005 priorities
<DaveB> variable names will have additional restrictions
<DaveB> no '-'
<Yoshio> agree
<DaveB> and no '.' says AndyS, I just found his mail to check
<Yoshio> I'm happy
<Yoshio> anyway, thanks for clarifying
<Yoshio> I'll wait for the new syntax
<DaveB> ha, I found a test bug
<DaveB> and I agree with ARQ, so it's not my bug
<DaveB> ex11.2.3.6_0.rq ex11.2.3.6_0.n3
<DaveB> answer should be
<DaveB> -----------------------------------------
<DaveB> | name | mbox |
<DaveB> =========================================
<DaveB> | "Alice" | <mailto:alice@work.example> |
<DaveB> -----------------------------------------
<DaveB> result file says rs:value "alice@work.example" ;
<DaveB> it's a uri, not a string
<DaveB> RRSAgent: please leave us