IRC log of dawg on 2005-08-09

Timestamps are in UTC.

14:29:35 [RRSAgent]

RRSAgent has joined #dawg

14:29:35 [RRSAgent]

logging to http://www.w3.org/2005/08/09-dawg-irc

14:30:03 [JanneS]

thanks

14:30:08 [DanC]

Zakim, take up agendum 1

14:30:08 [Zakim]

agendum 1. "Convene, take roll, review records and agenda" taken up [from DanC]

14:30:25 [DanC]

Zakim, who's on the phone?

14:30:25 [Zakim]

On the phone I see JanneS, ILRT, DanC

14:30:26 [Zakim]

ILRT has DaveB, AndyS, ericP

14:30:48 [Zakim]

+??P6

14:31:03 [SteveH]

Zakim, ??P6 is SteveH

14:31:03 [Zakim]

+SteveH; got it

14:31:16 [Zakim]

+Kendall_Clark

14:31:39 [JanneS]

http://lists.w3.org/Archives/Public/public-rdf-dawg/2005JulSep/0175.html

14:31:42 [JanneS]

last week's minutes are there

14:31:44 [DanC]

-> http://lists.w3.org/Archives/Public/public-rdf-dawg/2005JulSep/0175.html Minutes of the DAWG telecon 2005-08-02

14:32:28 [kendall]

I sent regrets for 23 August

14:33:01 [ericP]

zakim, who is here?

14:33:01 [Zakim]

On the phone I see JanneS, ILRT, DanC (muted), SteveH, Kendall_Clark

14:33:02 [Zakim]

ILRT has DaveB, AndyS, ericP

14:33:03 [Zakim]

On IRC I see RRSAgent, kendall, JanneS, DaveB, Zakim, AndyS, SteveH, ericP, DanC

14:33:49 [DanC]

KC, JanneS at risk next week.

14:33:51 [kendall]

but i'm otherwise not at risk

14:34:13 [DanC]

Zakim, pick a scribe

14:34:13 [Zakim]

Not knowing who is chairing or who scribed recently, I propose SteveH

14:34:22 [JanneS]

Howard and Lee have sent regrets for 16-Aug

14:34:24 [DanC]

Zakim, pick a scribe

14:34:24 [Zakim]

Not knowing who is chairing or who scribed recently, I propose SteveH

14:34:27 [DanC]

Zakim, pick a scribe

14:34:28 [Zakim]

Not knowing who is chairing or who scribed recently, I propose SteveH

14:34:34 [DanC]

Zakim, pick a scribe

14:34:34 [Zakim]

Not knowing who is chairing or who scribed recently, I propose JanneS

14:34:36 [DanC]

Zakim, pick a scribe

14:34:36 [Zakim]

Not knowing who is chairing or who scribed recently, I propose ericP

14:34:43 [DanC]

Zakim, pick a scribe

14:34:43 [Zakim]

Not knowing who is chairing or who scribed recently, I propose Kendall_Clark

14:35:09 [DanC]

scribe: Steve or else EricP

14:35:47 [DanC]

agenda + WSDL 2 last call

14:35:53 [kendall]

zakim, mute me

14:35:53 [Zakim]

Kendall_Clark should now be muted

14:36:05 [DanC]

Zakim next item

14:36:09 [DanC]

Zakim, next item

14:36:09 [Zakim]

agendum 2. "comment "Query forms should be resources, not operations"" taken up

14:36:11 [kendall]

zakim, unmute me

14:36:11 [Zakim]

Kendall_Clark should no longer be muted

14:38:36 [Zakim]

+[IBMCambridge]

14:38:46 [JanneS]

discussion about http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2005Aug/0036

14:38:53 [DaveB]

markB's original comment http://lists.w3.org/Archives/Public/public-rdf-dawg/2005JulSep/0227.html

14:39:20 [EliasT]

EliasT has joined #dawg

14:39:43 [JanneS]

KendallC: splitting the operation into GET uri requires e.g. more intelligent clients to separate the operation from the expression to the protocol request

14:40:26 [ericP]

q+ to suggest that Mark get a round to produce a use case

14:40:55 [EliasT]

Zakim, who's on the phone?

14:40:55 [Zakim]

On the phone I see JanneS, ILRT, DanC, SteveH, Kendall_Clark, [IBMCambridge]

14:40:57 [Zakim]

ILRT has DaveB, AndyS, ericP

14:41:03 [EliasT]

Zakim, IBMCambridge is EliasT

14:41:03 [Zakim]

+EliasT; got it

14:41:16 [ericP]

q-

14:41:28 [JanneS]

KendallC: will formulate a response and having it circulated via WG before officially getting back to Mark

14:42:09 [kendall]

ACTION KendallC: write another draft response to Mark Baker's comment and send to WG for possible discussion

14:42:21 [DanC]

Zakim, next item

14:42:21 [Zakim]

agendum 3. "toward Protocol last call" taken up

14:42:46 [DanC]

1.57 http://www.w3.org/2001/sw/DataAccess/proto-wd/

14:43:39 [AndyS]

(There are two k examples - first has a syntax error)

14:44:04 [AndyS]

(OK - I see why now)

14:44:11 [kendall]

yes, the first k. has an intentional...gotcha

14:44:40 [DaveB]

B.d

14:44:44 [DaveB]

pls add sub-section nos :)

14:46:03 [JanneS]

the subsection id would be 2.B.d I guess

14:50:15 [DaveB]

nit - the sparql query results in the refernece section needs date, dated url, a bit of expansion

14:51:30 [kendall]

DaveB: yep, thx, all those refs need to be fixed up

14:51:37 [DaveB]

IIRC you'll have to change (example) domains from my.other.example to something.example.{com|org|net)

14:52:00 [DanC]

no, .example is a TLD

14:55:29 [kendall]

yeah, i thought it was foo.example, for example hosts, but now I don't know :>

14:56:45 [AndyS]

http://en.wikipedia.org/wiki/.example

14:57:39 [DanC]

ACTION KC: abstract/abbreviate http trace examples

14:57:47 [EliasT]

AndyS, I think Joseki3 can only take on --data parameter.

14:57:51 [DanC]

ACTION DanC: extract and machine-check examples

14:57:55 [EliasT]

s/on/one

14:58:11 [AndyS]

I find that people don't know any % encoding (cos browsers don't require it)

14:58:13 [DanC]

. ACTION SteveH: elaborate "CONSTRUCT with content negotiation" into a test case

14:58:19 [kendall]

Ah, so .example is safe?

14:58:33 [AndyS]

Elias - --data to Joseki3? Client?

14:58:48 [EliasT]

yes. sorry. arq.

14:59:23 [kendall]

ACTION KendallC: fix d. ASK with simple RDF dataset...XML is wrong.

14:59:32 [AndyS]

Shoudl support multiple - I'll check (which version?)

15:00:02 [EliasT]

Last week CVS Head.

15:00:41 [EliasT]

I might be wrong. I'll double check if you like.

15:00:48 [AndyS]

Ah - THAT version :-) So the ARQ client only accepts one --data? i.e. one default-uri= in protocol request?

15:00:54 [EliasT]

yes.

15:01:29 [AndyS]

I'll check - you're likely to be right - plenty of chances for me to have got it wrong :-(

15:02:00 [AndyS]

arq.sparql shoudl accept multiple --data and these should get into the protocol request

15:02:13 [AndyS]

and you have it in the IRC log now :-)

15:02:23 [kendall]

who's scribing today?

15:02:24 [EliasT]

:-)

15:02:29 [JanneS]

I'm trying

15:02:30 [DanC]

ACTION SteveH: elaborate "CONSTRUCT with content negotiation" into a test case [continued/elaborated]

15:02:30 [EliasT]

JanneS I believe.

15:02:49 [EliasT]

KendallC, I can help you with the refactoring of the examples.

15:02:56 [SteveH]

note to self, break results into good and acceptable

15:03:02 [kendall]

JanneS: can you make sure all of the comments from people re: protocol spec make it into the minutes? My irc client isn't scrolling back today, and I don't want to miss any feedback.

15:03:22 [kendall]

Elias: cool. i'd appreciate the help.

15:03:23 [JanneS]

ok, I'm doing my best.

15:03:49 [JanneS]

raw data is getting logged to http://www.w3.org/2005/08/09-dawg-irc already

15:03:58 [kendall]

oh, cool. that'll work.

15:05:35 [EliasT]

kendall, I guess we're going to need a dataset representation.

15:06:04 [EliasT]

kendall, this might do http://www.w3.org/2001/sw/DataAccess/tests/test-query.n3

15:06:58 [JanneS]

KendallC: 2 things still so decide before LC: 1. how to return faults in SOAP/WSDL

15:08:06 [AndyS]

"Figure 1.3 WSDL 2.0 fragment" ==> does not agree with WSDL

15:11:21 [JanneS]

KendallC: AndyS also commented on the fact that a server even supporting SPARQL extensions may have to return fault as the SPARQL query expression may not be valid w.r.t w3c rec

15:12:22 [JanneS]

DanC: such extensions can only be semantical but not syntactical. Syntactical non-conformance with the sparql grammar MUST be faults

15:12:49 [kendall]

I know I don't want to do *that*.

15:15:11 [JanneS]

DanC: we can endorse the design where a SPARQL service may support other languages than SPARQL - or we may limit what a SPARQL service may do. (I hope I got this right)

15:17:06 [JanneS]

KendallC: from WSDL perspective the query is only a string.

15:17:16 [AndyS]

<xs:element minOccurs="1" maxOccurs="1" name="sparql-query" type="xs:string">

15:17:18 [JanneS]

DanC: did you define a SPARQL specific type in WSDL?

15:17:22 [JanneS]

KendallC: nope

15:17:39 [DanC]

"called st:query in Figure 1.0, composed of two further parts: one SPARQL query string"

15:20:07 [DanC]

PROPOSED: a SPARQL service _must_ return MalformedQuery for queries that are not SPARQL Query Strings

15:20:15 [JanneS]

AndyS: we could leave the non-sparql-query response unspecified

15:21:59 [kendall]

AndyS: i'm actually fine with it too...

15:22:02 [JanneS]

DaveB: don't know how wsdl supports this but one could enforce the behaviour if the expression is SPARQL.

15:23:06 [kendall]

in wsdl you could just add another operation to the interface

15:24:05 [SteveH]

seems like a circular argument. if the query processor is not SPARQL then it doesnt follow the spec, and it fails the may... *shrug*

15:24:55 [kendall]

a superset of foo isn't a foo?

15:25:29 [JanneS]

is anyone planning to do a subset of foo? :-/

15:25:51 [kendall]

heh!

15:27:07 [DanC]

(interesting question... what's the state of the art in interface evolution? are there best practices?)

15:29:17 [kendall]

it's tricky

15:29:33 [kendall]

though the conversatives seemed to have a small majority here, today

15:30:55 [JanneS]

DanC: conservative approach might be put on the table now... what do you all think?

15:31:57 [kendall]

Hmm, I'm switching my vote to the conservative side (for the first time, ever, I will have voted conservative!)

15:34:05 [JanneS]

no lessons to learn from HTTP 1.0 -> 1.1 :_(

15:37:07 [DanC]

PROPOSED: a SPARQL service _must_ return MalformedQuery for queries that are not SPARQL Query Strings

15:37:39 [DanC]

RESOLVED, W3C/EricP and UMD/KC abstaining

15:40:53 [DanC]

ACTION EricP: review protocol document 1.58

15:41:20 [DanC]

ACTION SteveH: review proto-wd after Friday signal from KC, hopefully in time for monday's agenda or tuesday's telcon

15:41:40 [kendall]

eek, sorry to eat so much agenda

15:41:47 [DanC]

Zakim, next item

15:41:47 [Zakim]

agendum 4. "comment "Security Considerations"" taken up

15:42:27 [kendall]

3. Policy Considerations

15:42:34 [kendall]

with security and privacy sections

15:42:48 [DaveB]

there's a section in xml res, in ietf registration

15:43:18 [DaveB]

in http://www.w3.org/TR/2005/WD-rdf-sparql-XMLres-20050801/#mime

15:43:33 [DanC]

ACTION EricP: add "see also protocol security considerations" to rq23

15:44:19 [ericP]

http://www.w3.org/TR/xquery/#N1805C

15:44:27 [ericP]

http://www.ietf.org/rfc/rfc1738.txt

15:44:28 [kendall]

ah, interesting analogue

15:44:29 [JanneS]

KendallC: I have included the denial-of-service aspects of the protocol in the editor's draft

15:44:54 [ericP]

http://www.w3.org/2001/sw/DataAccess/rq23/#QTSyn

15:45:32 [DaveB]

(did the query lang mime type section get decided? i drafted a security considerations section in that)

15:45:33 [DanC]

ACTION 8 = EricP: add "security considerations" to rq23, incl "see also protocol security considerations"

15:45:43 [DaveB]

in http://www.w3.org/2001/sw/DataAccess/rq23/mime.txt

15:46:13 [DanC]

(that was continued without discussion, daveb; we're not done with the queryMimeType issue... er... or maybe we did decide but there are actions pending)

15:47:13 [kendall]

I'd be happy to have that discussion in one doc or the other, with a link from the one where it isn't, to the one where it is.

15:47:20 [SteveH]

q+ to talk about multiple froms

15:47:33 [DanC]

ack SteveH

15:47:33 [Zakim]

SteveH, you wanted to talk about multiple froms

15:48:34 [JanneS]

SteveH: multiple HTTP requests need to be flagged in case of multiple FROMs allowed (IETF requirement - ref?)

15:49:18 [AndyS]

+ FROM NAMED

15:49:24 [DanC]

Zakim, next item

15:49:24 [Zakim]

agendum 5. ""SPARQL Query Results XML Format: XML 1.1" comment" taken up

15:49:39 [kendall]

yep, same deal with from & from named

15:49:50 [JanneS]

http://www.w3.org/mid/1123099748.20865.18.camel@localhost

15:50:30 [DanC]

[[

15:50:31 [DanC]

For

15:50:31 [DanC]

XML 1.1 documents, individuals can use the XML Schema as indicated in

15:50:31 [DanC]

[1].

15:50:31 [DanC]

Philippe

15:50:32 [DanC]

[1] http://www.w3.org/TR/2005/NOTE-xml11schema10-20050511/

15:50:34 [DanC]

]]

15:51:00 [kendall]

Dan has become such a logical positivist! :>

15:51:13 [kendall]

the "C" in "DanC" stands for "Carnap"!

15:52:03 [DanC]

ACTION DaveB: see if PLH would be satisfied with s/can/may/ in what he suggeste

15:52:11 [DanC]

Zakim, next item

15:52:11 [Zakim]

agendum 6. ""SPARQL Results Format and Unbound Variables" comment" taken up

15:52:33 [JanneS]

http://www.w3.org/mid/42F4CEEB.5090306@umd.edu

15:52:40 [ericP]

-> http://www.w3.org/2001/sw/DataAccess/rq23/#security Security Considerations

15:53:23 [ericP]

i have written a Security Considerations section and committed it in rq23 1.460

15:53:36 [JanneS]

I'll note your action item immediately done :)

15:53:54 [ericP]

who's a good boy?

15:54:27 [JanneS]

did you get your gameboy back yet? You should...

15:54:38 [kendall]

bah

15:55:33 [kendall]

I can't hear DaveB well on the phone today. Too bad.

15:56:52 [DanC]

DaveB: the present design supports consumers that want just one way to search for the binding of a variable

15:57:51 [kendall]

I think the suggestion was to change <variable...><unbound/></variable> to <variable name=""/>...

15:58:03 [SteveH]

DanC, yes, i have but ive not run figures on it

15:59:19 [kendall]

er, s/variable/binding/

16:00:21 [kendall]

giving binding an unbound attribute doesn't seem to be too hard to check xpath'ly

16:00:43 [DanC]

ADJOURN.

16:00:44 [Zakim]

-SteveH

16:00:45 [Zakim]

-EliasT

16:00:48 [Zakim]

-ILRT

16:00:50 [Zakim]

-Kendall_Clark

16:01:06 [DanC]

RRSAgent, make logs world-access

16:01:38 [Zakim]

-DanC

16:01:40 [Zakim]

-JanneS

16:01:41 [Zakim]

SW_DAWG()10:30AM has ended

16:01:42 [Zakim]

Attendees were JanneS, DaveB, AndyS, ericP, DanC, SteveH, Kendall_Clark, EliasT

16:02:11 [DanC]

kendall, you might send mail to the WG showing the not-too-hard XPaths. or ask somebody at UMD to supplement the comment with them

16:03:25 [SteveH]

the thing about bandwidth v's easy of XSLT, is that writing a hard XSLT hurts once, whereas your mobile device taking twice as long to get the results hurts for every query

16:04:13 [SteveH]

gzip helps, but its not ideal

16:16:40 [DanC]

good point, SteveH ... but... writing a hard XSLT hurts once for every new developer that picks up SPARQL

16:34:42 [ericP]

it's hard to compete with the kind of pain you can get from XSLT

17:50:26 [SteveH]

DanC, not neccesarily, as most will hack pre-written XSLT sheets, or use DOM/SAX/whatever

18:20:26 [Zakim]

Zakim has left #dawg

18:29:46 [DanC]

RRSAgent, bye

18:29:46 [RRSAgent]

I see 9 open action items saved in http://www.w3.org/2005/08/09-dawg-actions.rdf :

18:29:46 [RRSAgent]

ACTION: KendallC to write another draft response to Mark Baker's comment and send to WG for possible discussion [1]

18:29:46 [RRSAgent]

recorded in http://www.w3.org/2005/08/09-dawg-irc#T14-42-09

18:29:46 [RRSAgent]

ACTION: KC to abstract/abbreviate http trace examples [2]

18:29:46 [RRSAgent]

recorded in http://www.w3.org/2005/08/09-dawg-irc#T14-57-39

18:29:46 [RRSAgent]

ACTION: DanC to extract and machine-check examples [3]

18:29:46 [RRSAgent]

recorded in http://www.w3.org/2005/08/09-dawg-irc#T14-57-51

18:29:46 [RRSAgent]

ACTION: KendallC to fix d. ASK with simple RDF dataset...XML is wrong. [4]

18:29:46 [RRSAgent]

recorded in http://www.w3.org/2005/08/09-dawg-irc#T14-59-23

18:29:46 [RRSAgent]

ACTION: SteveH to elaborate "CONSTRUCT with content negotiation" into a test case [continued/elaborated] [5]

18:29:46 [RRSAgent]

recorded in http://www.w3.org/2005/08/09-dawg-irc#T15-02-30

18:29:46 [RRSAgent]

ACTION: EricP to review protocol document 1.58 [6]

18:29:46 [RRSAgent]

recorded in http://www.w3.org/2005/08/09-dawg-irc#T15-40-53

18:29:46 [RRSAgent]

ACTION: SteveH to review proto-wd after Friday signal from KC, hopefully in time for monday's agenda or tuesday's telcon [7]

18:29:46 [RRSAgent]

recorded in http://www.w3.org/2005/08/09-dawg-irc#T15-41-20

18:29:46 [RRSAgent]

ACTION: EricP: add "security considerations" to rq23, incl "see also protocol security considerations" [8]

18:29:46 [RRSAgent]

recorded in http://www.w3.org/2005/08/09-dawg-irc#T15-43-33

18:29:46 [RRSAgent]

ACTION: DaveB to see if PLH would be satisfied with s/can/may/ in what he suggeste [9]

18:29:46 [RRSAgent]

recorded in http://www.w3.org/2005/08/09-dawg-irc#T15-52-03