Minutes of RDF DAWG telecon Tuesday 2004-06-15 for review from Farrukh Najmi on 2004-06-16 (public-rdf-dawg@w3.org from April to June 2004) (original) (raw)

Please look for "?" for areas where I was not sure of something. Thanks to all for their help during my first lame attempt at being scribe.

Minutes of RDF DAWG telecon Tuesday 2004-06-15 for review

Agenda http://lists.w3.org/Archives/Public/public-rdf-dawg/2004AprJun/0625.html

IRC log http://www.w3.org/2004/06/15-dawg-irc

ACTIONS ALL Register for F2F2 at http://www.w3.org/2002/09/wbs/35463/DAWGf2f2/ Kendall to consider the larger comments from Brian and Yoshio post-publication Tom to follow up on -comments mailing list on "How Many" proposed requirement. Simon to post to list and explain how a variable bind with zero variables is equivalent to boolean queries SteveH to add ericP's disjunctions tests to test cases doc Action: JimH elaborate his proposed use case (RDFS/OWL related) in story form Kendall to strike 3.4a from doc Kendall to move 3.4's status to ACCEPTED. DanC to rephrase 3.8 Bookmarkable Queries with reference to protocol AndyS to write email clarifying this?? JimH to send mail describing the history of remote KB query, including any mention of Denial of Service problems that may be relevant. EricP to chair in DanC absence in WG 6/22 meeting. DanC traveling next week.

Actions From Previous meeting: EricP to send announcement of UC&R to rdfs-rules list Kendall to dispose of "3.3a Extensible Value Testing" variant EricP to send an example to the list (as a test case) RobS to tweak it that it doesn't work ?? Simon draft 2 design proposals DanC to rephrase 3.8 with reference to protocol AndyS to email text for a requirement on streamable results

RESOLVED to strike 3.4a and ACCEPT 3.4

Convene, take roll, review record, agenda

RDF Data Access Working Group http://www.w3.org/2001/sw/DataAccess/ Tuesday 2004-06-15 14:30 UTC* Zakim Bridge +1.617.761.6200, conference 7333 ("RDFD") http://www.w3.org/Guide/1998/08/teleconference-calendar#s_1343 supplementary chat: irc://irc.w3.org:6665/dawg log to appear: http://www.w3.org/2004/06/08-dawg-irc

Scribe: Farrukh Najmi

(All, Please check roll call for omissions/errors??)

Present: Andy Seaborne Alberto Reggiori Dan Connolly (chair) Dave Beckett Dirk Colaert Eric Prud'hommeaux Farrukh Najmi (scribe) James Hendler Jos De Roo Kendall Clark Kevin Wilkinson Pat Hayes Rob Shearer Simon Raboczi Steve Harris Yoshio Fukushige Janne Saarela Tom Adams Regrets: Howard Katz

Minutes of RDF DAWG telcon 2004-06-01 for review http://lists.w3.org/Archives/Public/public-rdf-dawg/2004AprJun/0503.html

Minutes of RDF DAWG telecon Tuesday 2004-06-08 for review http://lists.w3.org/Archives/Public/public-rdf-dawg/2004AprJun/0547.html as ammended Wed, 09 Jun 2004 08:14:00 -0500 by Connolly

APPROVED as a true record of the meeting Next meeting: 2004-06-22 14:30 UTC Scribe: ??

comments on the agenda?

Ken: Need to review additional new use cases

ftf update

ALL F2F2 registration at http://www.w3.org/2002/09/wbs/35463/DAWGf2f2/

RobS: 7 people have replied so far DanC: Use reg form for ride coordination instead of WG list Not many people coming day of meeting

ACTION: ALL Register for F2F2 at http://www.w3.org/2002/09/wbs/35463/DAWGf2f2/ RESOLVED: Start 9AM on day 1 of f2f

Use Cases and Requirements Doc

ACTION DaveB: write up this educational metadata UC

DONE: http://lists.w3.org/Archives/Public/public-rdf-dawg/2004AprJun/0628.html

Several members: agreed that free text search would be a useful addition to Requirements. It is motivated by above new use case.

??: Should we add this as a new use case

Several members: speak in favor of adding this new use case. None against.

DanC: Editor has been advised to include

DaveB: yay

--Discussion on: New use case - RDFS/OWL related

http://lists.w3.org/Archives/Public/public-rdf-dawg/2004AprJun/0529.html

RobS: concerned about endorsing approach to OWL that doesn't work in general Yoshio:What is the relationship to the "tell me about" query issue DanC: Understand that the use case deals with Vanilla Graph queries and not "Tell Me About" queries AndyS: Agreed with DanC

Action: JimH elaborate his proposed use case in story form

ACTION: Kendall: consider the larger comments from Brian and Yoshio post-publication

Continues

ACTION EricP to send announcement of UC&R to rdfs-rules list

DONE

ACTION Kendall to dispose of "3.3a Extensible Value Testing" variant

DONE

ACTION Kendall effect "RESOLVED to adopt 3.9 as an objective" (not recorded)

DONE

--Discussion on External Comments

EricP: pointer to the -comments list: http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2004Jun/thread.html

DanC: Anyone read incoming comments on UC&R?

alberto: it seems Jim use-cased relates to DaveB about "give me useful stuff about X" and the RDF/S and/or thesaurus API - see http://lists.w3.org/Archives/Public/public-rdf-dawg/2004JanMar/0063.html

DaveB: Thinking about my new UC, I think it also motivates (at least allowing) RDFS inferencing

DanC: Bjoern Hoehrmann writes about How Many" requirements. This one scares me. http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2004Jun/0004.html

tom: I think a "count" is very important. Simon: Can you comment? SimonR: tom: I agree, but I'm listening to other people for a moment about it....

DanC: Inclined to say NO to "How Many" proposed requirement

SimonR: One thing we could do is support aggregate functions that are functions of a result

AndyS: I think we're talking about counts over the result sets.

ericP: +1

SimonR: Typical aggregate functions: count(), average(), minimum(), etc. SteveH: +1 to Andy

Kevin: This is related to limits. If we did limits then not a big extension to do count

kendall: +1 on danc's type/token distinction

DanC: Next step is for someone to propose what to do about this

tom: Can we not bind the value of a count to a variable, feed this back into a query? We certainly use this feature a lot.

JosD: You cannot avoid the cases where count is useful

Yoshio: average() ? then variance()... I think count () is enough SteveH: Yoshio: i agree SimonR: Yoshio: The question of which aggregate functions we support is a separate issue, but there should be some generic method of including them. Yoshio: I agree to SimonR

DanC: Can someone write back to sender that this is interesting point. Avoid any commitments on behalf of WG. tom: Willing to write back

ACTION: Tom to follow up on -comments mailing list on "How Many" proposed requirement.

--Discussion on ?? (lost this part of conversation almost entirely)

SimonR: DanC, Is this content selection for mobile devices draft based on CC/PP?

SteveH: Maybe we can get them to write what the query might look like

SimonR?: Would be nice to have a liaison

DanC: acceptable to not liaison with DI, but it would be nice to do so.

--Discussion on ??

rob_: boolean or "grounded" queries ericP: count(*) > 0 is a true or false answer ?: Want to be able to form a query that return boolean values afs: Or "LIMIT 1" - allows the query to stop as soon as one thing found. ?: Special case of zero variables returning boolean value is a special case of variable bindings

ACTION Simon to write to the list and explain how a variable bind with zero variables is equivalent to boolean queries.

Review test materials (ACTION SteveH)

Testcase queries temporarily available at: http://www.ecs.soton.ac.uk/~swh/TestSketchCases/ ACTION: SteveH continues

DanC: Had discussion of datatype queries, somewhat tangential but did uncover some interesting impl experience

DanC: EricP has written up a couple of tests http://lists.w3.org/Archives/Public/public-rdf-dawg/2004AprJun/0615.html http://lists.w3.org/Archives/Public/public-rdf-dawg/2004AprJun/0598.html http://lists.w3.org/Archives/Public/public-rdf-dawg/2004AprJun/0630.html

ACTION: SteveH to add ericP's disjunctions tests to test cases doc

Requirements

-- 3.4 Subgraph Results EricP to send an example to the list (as a test case) RobS to tweak it that it doesn't work ??

DanC: Eric gave an example of subgraph results Yoshio: I wonder if we can live without the "sub" but just "graph results"

ACTION: JimH to elaborate his use case.

RobS: Does not explain how it matches either

RobS: Comes down to last 4 words of sub-graph results.

Kendall: Already stricken. http://www.w3.org/2001/sw/DataAccess/UseCases#r3.4

DanC: Anyone object to adopting r3.4 Simon,RobS, Yoshio: abstain

ACTION: Kendall to strike 3.4a from doc and move 3.4's status to ACCEPTED.

-- 3.5 Local Queries

ACTION: BryanT to draft possible new requirement on independence of protocol and query language.

-- 3.6 Optional Match

ACTION Simon draft 2 design proposals

?: How to format the display of sample results in tabular form. Kendall: Prefer ASCII format Several members: prefer ASCII

3.8 Bookmarkable Queries

ACTION DanC to rephrase 3.8 with reference to protocol

DanC: No update. ACTION continues

-- 3.9 Bandwidth-efficient Protocol

AndyS to email text for a requirement on streamable results

DanC: round streamable doodad http://lists.w3.org/Archives/Public/public-rdf-dawg/2004AprJun/0566.html DaveB: my reply http://lists.w3.org/Archives/Public/public-rdf-dawg/2004AprJun/0569.html

AndyS: Would rather not have optional addition DaveB: "When returning multiple results there must be a form of the protocol that allows streaming of results. In this form, the results must be returned to the client so that all the data in one result are available before all the data for the next result." +"This does not require an ordering of results"

Several people for adding this requirements.

RobS: Raises concern about adding this requirement RobS: Not worded in terms that client is asking for streaming of results RobS: When return multiple results is irrelevant RobS: The fact that such a protocol exists does not mean the server is using it DanC: Do you have a technical reason for not supporting this? RobS: The words in the requirement are not requiring that server stream results

ACTION: Andy will write email clarifying this

JimH: A lot of things that do remote SQL server designed for knowledge base queries instead of db queries often assume that you have a lot of results JimH: points out that many remote KB query protocols don't return results but, rather, a number and a pointer, so that client could "control" the returning of results. JimH mentions OKBC, Parka, SQL, other stuff Kendal: We need to ack that there is significant support to add this use case

ACTION: JimH to send mail describing the history of remote KB query, including any mention of Denial of Service problems that may be relevant.

ACTION: EricP to chair in DanC absence. DanC traveling next week. 15:55:09 [Zakim]

-- Regards, Farrukh

**** BEGIN LOGGING AT Tue Jun 8 07:58:01 2004

Jun 08 07:58:01 --> You are now talking on #dawg Jun 08 07:58:26 --- DanC has changed the topic to: RDF Data Access WG 8 Jun http://www.w3.org/2001/sw/DataAccess/ Jun 08 09:07:52 --> SimonR (raboczi@138.130.248.69) has joined #dawg Jun 08 09:20:36 --> KevinW (Kevin.Wilk@192.6.19.124) has joined #dawg Jun 08 09:20:49 --> AndyS (AndyS@192.6.10.2) has joined #dawg Jun 08 09:23:44 --> howardk (howardk@24.207.8.174) has joined #dawg Jun 08 09:25:52 --> SteveH (swh@152.78.190.193) has joined #dawg Jun 08 09:26:52 --> FarrukhNajmi (najmi@192.18.98.63) has joined #dawg Jun 08 09:27:30 * howardk might have to leave early Jun 08 09:27:47 --- You're inviting ericP to #dawg (mcclure.w3.org) Jun 08 09:27:47 --- DanC invited ericP into channel #dawg. Jun 08 09:28:02 --> alberto (alberto@81.74.1.226) has joined #dawg Jun 08 09:28:06 --> Yoshio (fuku@18.29.0.64) has joined #dawg Jun 08 09:28:21 --> JosD (chatzilla@194.7.105.71) has joined #dawg Jun 08 09:29:48 --> kendall (Kendall@216.164.63.22) has joined #dawg Jun 08 09:30:02 --> DaveB (dajobe@137.222.34.229) has joined #dawg Jun 08 09:31:12 hello, all Jun 08 09:31:28 ohayoo yoshio-san Jun 08 09:31:31 hey Jun 08 09:31:36 Good evening! Jun 08 09:31:43 hello Jun 08 09:31:48 hola, que tal! Jun 08 09:31:54 Oh, kombanwa, howard-san :) Jun 08 09:32:05 kombanwa desu ka? bikkuri! Jun 08 09:32:27 Yes, it's 23:30 here Jun 08 09:32:50 :-) Jun 08 09:33:04 --> ericP (ericP@18.29.0.30) has joined #dawg Jun 08 09:33:31 a slow start. that's good, since I'm hung over Jun 08 09:34:00 hung over on a tuesday?! Howard the K indeed!! :> Jun 08 09:34:17 nah, i was just funnin' Jun 08 09:34:22 * kendall been celebrating the Republic Death Festival in the US non-stop since Saturday. -wink- Jun 08 09:34:27 --> rob_ (rob@66.146.191.148) has joined #dawg Jun 08 09:34:29 er, s/Republic/Republican/ Jun 08 09:34:33 you're supposed to say "is it tuesday already"? Jun 08 09:35:07 It's Tuesday? - no one told me Jun 08 09:35:31 i only new because my phone told me :) Jun 08 09:35:55 Zakim, who is here Jun 08 09:35:58 smart phone! Jun 08 09:36:11 "It must be Monday - the agenda arrived" Jun 08 09:36:17 I see no Zakim yet Jun 08 09:36:22 wow, tappie-tappie Jun 08 09:36:43 Zakim, wake up! Jun 08 09:36:52 Zakim speak! Jun 08 09:36:53 someone ping me when we get started? I'm hanging up. Jun 08 09:37:01 Zakim: Oh is it Tuesday? Jun 08 09:37:22 * alberto the keybord does not sound like an Apple ;) Jun 08 09:37:50 kendall: what is the Republican Death Festival?? a funnie? Jun 08 09:38:04 Zakim, who is on the phone Jun 08 09:38:08 howard: all the maudlin nonsense re: reagan's death. Jun 08 09:38:13 ah! Jun 08 09:38:21 i live in downtown DC, and it's driving me crazy Jun 08 09:38:37 i'm heading out of town thur night to avoid all the insane security over the funeral Jun 08 09:38:38 --- ericP invited Zakim into channel #dawg. Jun 08 09:38:38 --> Zakim (rrs-bridgg@18.29.0.30) has joined #dawg Jun 08 09:38:50 zakim, this is dawg Jun 08 09:38:50 ok, ericP; that matches SW_DAWG()10:30AM Jun 08 09:38:57 plus, if i share my true feelings about this event with my fellow citizens, i'm liable to be beaten :> Jun 08 09:39:01 zakim, who's here? Jun 08 09:39:01 On the phone I see ??P5, Kevin, Farrukh, ??P8, HowardK, ??P10, Yoshio, ??P12, ChrisR, PatH, ??P15, RobS, EricP Jun 08 09:39:04 On IRC I see rob_, ericP, DaveB, kendall, JosD, Yoshio, alberto, FarrukhNajmi, SteveH, howardk, AndyS, KevinW, SimonR, DanC Jun 08 09:39:15 tarring and feathering has a fine tradition in your country! Jun 08 09:39:24 'zactly Jun 08 09:39:31 +Kendall_Clark Jun 08 09:40:15 +JimH Jun 08 09:41:06 Zakim, ??P5 is Josd+DirkC Jun 08 09:41:06 +Josd+DirkC; got it Jun 08 09:41:28 Zakim, who is on the phone Jun 08 09:41:28 I don't understand 'who is on the phone', JosD Jun 08 09:41:38 zakim, who's here Jun 08 09:41:38 kendall, you need to end that query with '?' Jun 08 09:41:41 zakim, who's here? Jun 08 09:41:41 On the phone I see Josd+DirkC, Kevin, Farrukh, ??P8, HowardK, ??P10, Yoshio, ??P12, ChrisR, PatH, ??P15, RobS, EricP, Kendall_Clark, JimH Jun 08 09:41:44 On IRC I see rob_, ericP, DaveB, kendall, JosD, Yoshio, alberto, FarrukhNajmi, SteveH, howardk, AndyS, KevinW, SimonR, DanC Jun 08 09:41:57 zakim, Josd+DirkC has JosD, DirkC Jun 08 09:41:57 +JosD, DirkC; got it Jun 08 09:42:14 I'm one of the ??P but not sure which. Jun 08 09:42:20 oops! Jun 08 09:42:25 * DanC Zakim, call DanC-BOS Jun 08 09:42:25 * Zakim ok, DanC; the call is being made Jun 08 09:42:25 +DanC Jun 08 09:43:02 simon, andy, alberto Jun 08 09:43:19 simon, andy, alberto, daveb, steveh Jun 08 09:43:57 * DanC Zakim, who's talking? Jun 08 09:44:00 --> phayes (phayes@65.212.118.4) has joined #dawg Jun 08 09:44:07 * Zakim DanC, listening for 10 seconds I heard sound from the following: Josd+DirkC (20%), RobS (75%), DanC (49%) Jun 08 09:44:30 RESOLVED: http://lists.w3.org/Archives/Public/public-rdf-dawg/2004AprJun/0503.html are OK as a record of 1Jun Jun 08 09:44:34 am I humming or something? Jun 08 09:44:36 Minutes of RDF DAWG telcon 2004-06-01 APPROVED Jun 08 09:44:42 is that yr keyboard rob? Jun 08 09:44:50 no Jun 08 09:45:21 Zakim, pick a scribe Jun 08 09:45:21 Not knowing who is chairing or who scribed recently, I propose Farrukh Jun 08 09:46:01 next meeting june 15 scribe volunteer Farrukh Jun 08 09:46:28 Farrukh is scribe Jun 08 09:46:35 ==== 2. Status of July F2F Jun 08 09:46:47 Rob organizing logistics Jun 08 09:47:07 Happening Carlsbad CA July 14-15 Jun 08 09:47:50 Eric P has setup registration a couple of hours ago Jun 08 09:48:03 EricP to set up registration for F2F: DONE Jun 08 09:48:38 yay Jun 08 09:48:41 ==== 3. Publication of Use Cases and Requirements (UC&R) document Jun 08 09:50:30 . http://lists.w3.org/Archives/Public/www-rdf-interest/ Jun 08 09:50:54 www-rdf-rules ericp to announce Jun 08 09:50:59 ACTION: EricP to send announcement to rdfs-rules list Jun 08 09:51:25 klyne http://lists.w3.org/Archives/Public/public-rdf-dawg/2004AprJun/0521.html Jun 08 09:52:58 ... no immediate changes of words required; keep for later update Jun 08 09:53:43 ==== 4. Review requirements, evaluating designs Jun 08 09:54:15 http://www.w3.org/2001/sw/DataAccess/UseCases#req Jun 08 09:54:42 == 3.3 Extensible Value Testing Jun 08 09:56:32 ... Kendall's iTQL evaluation and Simon's reply Jun 08 09:56:58 finally found the registration for the carlsebad f2f: http://www.w3.org/2002/09/wbs/35463/DAWGf2f2/ Jun 08 09:57:12 it was quite a hunt, though Jun 08 09:57:59 == how to dispose of "3.3a Extensible Value Testing" variant? Jun 08 09:58:10 ericp, remind me to come back to that before we adjourn, please Jun 08 09:58:15 roger Jun 08 09:58:45 I seem not able to login to that f2f URL...something about wrong passwords... Jun 08 09:59:05 * ericP looks to see what he screwed up... Jun 08 09:59:46 it asked me for a p'word but i gave my usual one and got in -- fwiw Jun 08 09:59:48 rob's message http://lists.w3.org/Archives/Public/public-rdf-dawg/2004AprJun/0538.html Jun 08 09:59:49 re rob's action http://lists.w3.org/Archives/Public/public-rdf-dawg/2004AprJun/0538.html Jun 08 09:59:49 ACTION Kendall to dispose of "3.3a Extensible Value Testing" variant Jun 08 10:00:16 == 3.4 Subgraph Results Jun 08 10:00:16 I may have my passwords screwed up...forget it, thanks. Jun 08 10:00:52 q+ to ask what if the original queried graph is a "merged" graph? Jun 08 10:00:52 * Zakim sees Yoshio on the speaker queue Jun 08 10:01:57 ... DanC wants that people think up further for one week Jun 08 10:04:14 ACTION EricP send an example to the list (as a test case) Jun 08 10:04:41 ACTION RobS to tweak it that it doesn't work ?? Jun 08 10:06:11 * Yoshio wants more verbose irc (I can't hear you very well) ... Jun 08 10:07:09 I found rob's origianl words http://lists.w3.org/Archives/Public/public-rdf-dawg/2004AprJun/0372.html Jun 08 10:07:43 ... anothe week to think about it (only 2 prepared to vote) Jun 08 10:09:10 andy: want to leave the ability to answer queries with optional arcs. Jun 08 10:09:43 jim: with some queries, you don't knwo what type you'll get back? Jun 08 10:09:43 +1 optional arcs Jun 08 10:09:44 . Jun 08 10:10:41 who was speaking against 3.4a? Jun 08 10:11:00 robs Jun 08 10:11:14 3.4a Subgraph Results (variant) is worse for RobS Jun 08 10:11:38 == 3.5 Local Queries Jun 08 10:11:45 ACTION: BryanT to draft possible new requirement on independence of protocol and query language. Jun 08 10:12:07 re: 3.4 (a) I don't see the reason why we should restrict the answer to "subgraph"s Jun 08 10:12:26 BryanT's action continued Jun 08 10:12:45 == 3.6 Optional Match Jun 08 10:13:00 ... It must be possible to express a query that does not fail when some specified part of the query fails to match. Any such triples matched by this optional part, or variable bindings caused by this optional part, can be returned in the results, if requested. Jun 08 10:13:48 our RDQL does "mimic" optionals (simple OR/disjunction) using 'coma' operator - I posted about it sometime ago http://lists.w3.org/Archives/Public/public-rdf-dawg/2004AprJun/0317.html Jun 08 10:14:00 RobS: doesn't think that there is not enough justification for such a requirement Jun 08 10:14:13 s/not/ Jun 08 10:14:51 +q to ask if we can drop the 3.6 variant to simplify the doc and focus deliberations Jun 08 10:15:11 zakim, +q to ask if we can drop the 3.6 variant to simplify the doc and focus deliberations Jun 08 10:15:11 I don't understand '+q to ask if we can drop the 3.6 variant to simplify the doc and focus deliberations', kendall Jun 08 10:15:15 argh Jun 08 10:15:24 odd Jun 08 10:15:28 s/+q/q+/ Jun 08 10:15:31 * DanC q? Jun 08 10:15:31 * Zakim sees Yoshio on the speaker queue Jun 08 10:15:32 thx Jun 08 10:15:39 zakim, q+ to ask if we can drop the 3.6 variant to simplify the doc and focus deliberations Jun 08 10:15:39 I see Yoshio, kendall on the speaker queue Jun 08 10:15:44 q- Jun 08 10:15:44 * Zakim sees kendall on the speaker queue Jun 08 10:16:31 q+ to note that lots of QLs have optionals Jun 08 10:16:31 * Zakim sees kendall, ericP on the speaker queue Jun 08 10:17:21 ack yoshio Jun 08 10:17:21 * Zakim sees kendall, ericP on the speaker queue Jun 08 10:17:33 q+ Jun 08 10:17:33 * Zakim sees kendall, ericP, phayes on the speaker queue Jun 08 10:17:44 oops; sorry, yoshio. Jun 08 10:17:53 (he wanted to speak on 3.4) Jun 08 10:17:57 * kendall can barely hear Yoshio. Jun 08 10🔞07 Simon: pretty in defense of this requirement Jun 08 10🔞21 ack kendall Jun 08 10🔞21 kendall, you wanted to ask if we can drop the 3.6 variant to simplify the doc and focus deliberations Jun 08 10🔞23 * Zakim sees ericP, phayes on the speaker queue Jun 08 10🔞47 kendall: anyone wnat to drop the optional match variant? Jun 08 10🔞50 ack ericp Jun 08 10🔞50 ericP, you wanted to note that lots of QLs have optionals Jun 08 10🔞51 q- Jun 08 10🔞52 * Zakim sees phayes on the speaker queue Jun 08 10🔞53 * Zakim sees phayes on the speaker queue Jun 08 10🔞53 danc: let's wait a bit Jun 08 10:19:33 Kendall asking to drop the 3.6 variant wording "It must be possible to express a query with optional parts such that the query does not fail to match when one or more optional parts of the query fails to match. Any such triples matched by this optional part, or variable bindings caused by this optional part, can be returned in the results, if requested." Jun 08 10:19:47 * DanC q? Jun 08 10:19:48 * Zakim sees phayes on the speaker queue Jun 08 10:19:50 This (3.6) is the #1 request I get for additional features Jun 08 10:20:03 jim: in favor of keeping optionals in Jun 08 10:20:12 1+ Jun 08 10:20:44 JimH: some conditional being done as an optional Jun 08 10:20:59 ... we're doing lots of extra queries that i think we can do more easily with a single query with optionals Jun 08 10:22:02 q+ Jun 08 10:22:02 * Zakim sees phayes, kendall on the speaker queue Jun 08 10:22:04 robS: i'm not sure that i'm at abstaining vs. objecting Jun 08 10:22:15 ... not saying that it's not a good idea Jun 08 10:22:49 * DanC Zakim, who's talking? Jun 08 10:22:51 simon: i can provide two defs: Jun 08 10:22:54 DanC: ignore the results if the first one fails, but how long will it take to get the details worked out, 2..3 months Jun 08 10:22:57 ... SQL outer joins Jun 08 10:23:00 * Zakim DanC, listening for 10 seconds I heard sound from the following: ??P10 (14%), PatH (9%), ??P15 (32%) Jun 08 10:23:06 comment about optionas: '*' or '//' operators in XPath are so useful - some similar articaft fro DAWG would be handy - RDF is unbounded and it is nice like that; you can "not" expect to know which data/triples a service has - your rather run "optional matches" (wildcards) to guess out Jun 08 10:23:08 ... something products withing products Jun 08 10:23:21 ACTION Simon draft 2 design proposals Jun 08 10:23:28 -JimH Jun 08 10:23:31 s/optionas/optionals/ Jun 08 10:23:37 * DanC encourages simon to elaborate on his action here while it's in his head Jun 08 10:23:45 * DanC q? Jun 08 10:23:45 * Zakim sees phayes, kendall on the speaker queue Jun 08 10:23:49 thx Jun 08 10:23:54 howard: is optional the same as conditional? there is a conditional in XQuery Jun 08 10:23:55 * kendall trying to use the queue :> Jun 08 10:24:11 ack phayes Jun 08 10:24:11 * Zakim sees kendall on the speaker queue Jun 08 10:24:32 ack kendall Jun 08 10:24:32 * Zakim sees no one on the speaker queue Jun 08 10:24:33 PatH: form not be possible to extremely desirable Jun 08 10:24:47 DanC: that's an objective Jun 08 10:24:59 kendall: re eric's survey of impls, we have to think about the politics Jun 08 10:25:27 Kendall: concerned about existing practice/users Jun 08 10:25:34 I'm not convinced optionals are that widely impl. (multiple impls, multiple langs) Jun 08 10:25:37 ... if we tell folks to switch to dawg-ql, and it doesn't have optoionals, it's a harder sell. Jun 08 10:26:43 DaveB: that's an interesting claim, if true ;> Jun 08 10:27:11 DanC considers moving it to an objective Jun 08 10:27:58 AndyS: let's finish the requirements discussion Jun 08 10:28:21 andyS: many of the proposed options seem to push us towards a style of ql, ala relational vs. graph Jun 08 10:28:26 == 3.8 Bookmarkable Queries Jun 08 10:28:32 ... other reqs don't do that Jun 08 10:29:01 RobS: a query expressed as some text Jun 08 10:29:05 robs: would rather have a requirement that a query shoudl be expressed as text Jun 08 10:29:52 The statement in the requirement that "the query is not intended to be humanly readable" seems to clash with one of the design goals for any URI.... Jun 08 10:29:56 How about: req that query can be expressed as an HTTP GET? Jun 08 10:29:58 AndyS: 'GET'able by HTTP Jun 08 10:30:38 +1 DanC Jun 08 10:30:42 robS: someone speak in favour of this Jun 08 10:31:16 danc: our mission is to promote interoperability. would consider my time well spent to get this done Jun 08 10:31:22 RobS: this is talking about the protocol Jun 08 10:31:40 * DanC q? Jun 08 10:31:40 * Zakim sees no one on the speaker queue Jun 08 10:31:57 AndyS: bookmarable query request Jun 08 10:32:01 andys: we could say "a protocol request" Jun 08 10:32:11 andys said: a query request Jun 08 10:32:24 right - tx Jun 08 10:32:49 ACTION DanC: rephrase 3.8 with reference to protocol Jun 08 10:32:59 Not sure who to addres to: I assumed that query itself does not necessarily need to be expressed in URI. It could persists somewhere and a URI could locate it. Is this correct? Jun 08 10:33:18 FarrukhNajmi: i dont think so Jun 08 10:33:19 == 3.9 Bandwidth-efficient Protocol Jun 08 10:33:57 SteveH message http://lists.w3.org/Archives/Public/public-rdf-dawg/2004AprJun/0501.html Jun 08 10:34:23 * howardk has to depart at least momentarily ... Jun 08 10:34:34 Thanks SteveH. I'll take a look. Jun 08 10:34:43 The access protocol design shall address bandwidth utilization issues; that is, it shall allow for at least one result format that does not make excessive use of network bandwidth for a given collection of results. Jun 08 10:35:24 * DanC Zakim, who's on the phone? Jun 08 10:35:24 * Zakim sees on the phone: Josd+DirkC, Kevin, Farrukh, ??P8, HowardK, ??P10, Yoshio, ??P12, ChrisR, PatH, ??P15, RobS, EricP, Kendall_Clark, DanC Jun 08 10:35:26 * Zakim Josd+DirkC has JosD, DirkC Jun 08 10:35:33 appeals to about 5 people Jun 08 10:35:44 6 or 7, maybe Jun 08 10:36:09 oops.. (lost my count) Jun 08 10:36:21 5 +/- 3 ;-) Jun 08 10:36:31 :-) Jun 08 10:36:43 simon: bandwidth-efficient protocol is hard to measure Jun 08 10:36:50 danc: +1 Jun 08 10:36:50 Hmm, re: 3.8 it sounds like a design objective at least to me(cf. 4.4) Jun 08 10:36:59 robs: let's keep in mind... Jun 08 10:37:08 RobS: it's more like a "let's keep in mind" Jun 08 10:37:18 steve: it is quantifiable Jun 08 10:37:25 ... which is then an objective Jun 08 10:37:46 +1 to 3.8 => Design Objective (and add latency as a DO) Jun 08 10:38:12 +1 Jun 08 10:38:16 if the protocol, say, did gzip or something, it'd meet this? Jun 08 10:38:29 daveb: presumably Jun 08 10:39:02 DanC: does anybody have customers with this requirement? Jun 08 10:39:25 ??: we do Jun 08 10:39:35 simon: we have features that won't work without efficiency Jun 08 10:39:57 s/??/simon Jun 08 10:40:42 ... could support latancy as a requirement and bandwidth as a DO Jun 08 10:41:00 ... the format in which.. must not preclude seriallization Jun 08 10:41:44 Don't we mean "streamable" not "serialize" RDF/XML is a serialization that is not streamable Jun 08 10:41:49 the only thing in here I see expressed in requirement form is "streamable" or "serializable" results Jun 08 10:42:24 DanC: people have argued that it is not testable Jun 08 10:42:38 (I think we mean "serialized processing of results") Jun 08 10:42:49 simon: we can test streamability Jun 08 10:43:03 Rob's general point about latency might be (?) that it must be possible for latency to be bounded independent of the form of the query? Jun 08 10:43:09 confirmed by AndyS Jun 08 10:43:11 andys: that would me you could process the first result before ... Jun 08 10:43:14 sees 3.9 related to 3.10a as protocol issues Jun 08 10:43:56 +1 streamability Jun 08 10:44:03 (also since I've already implemented that) Jun 08 10:44:07 ACTION AndyS: propose a steamabilyt requirement Jun 08 10:44:23 Action AndyS to email text for a requirement on streamable results Jun 08 10:44:23 ACTION AndyS to change wording wrt streamability Jun 08 10:44:40 [several in faaavor of moving 3.9 to a DO] Jun 08 10:44:49 and the vote... Jun 08 10:44:53 no objections Jun 08 10:45:01 no abstentions Jun 08 10:45:16 RESOLVED to move 3.9 to DOs Jun 08 10:45:19 RESOLVED to adopt 3.9 as an objective Jun 08 10:45:46 == 3.10 Result Limits Jun 08 10:46:02 It must be possible to to specify an upper bound on the number of query results returned. Jun 08 10:47:05 Andys: RDQL does not have a limit. Joseki provides limits. Jun 08 10:47:05 PatH: it could mean 2 things Jun 08 10:47:34 ... I don't want more than 10 answers Jun 08 10:47:49 ... and the next after that Jun 08 10:47:59 ... or just 10 period Jun 08 10:48:07 q+ to say that the SQL world was happy to hear "10 answers, don't know if there are more" Jun 08 10:48:07 * Zakim sees ericP on the speaker queue Jun 08 10:48:34 * DanC q? Jun 08 10:48:34 * Zakim sees ericP on the speaker queue Jun 08 10:49:54 andyS: seen this to prevent error conditions Jun 08 10:50:12 EricP: SQL has it -- RobS: other things that we don't have Jun 08 10:50:17 Suggest min req is to specify upper limit on results plus indicator that there are more results Jun 08 10:50:19 a query may want just 10 results; such as making a top-ten news feed, for eexample Jun 08 10:51:01 wondering about ordering now....which 10? first 10? Jun 08 10:51:25 to get "next 10" is very common with document retrieval Jun 08 10:51:26 * DanC q? Jun 08 10:51:26 * Zakim sees ericP on the speaker queue Jun 08 10:51:28 q+ to say that asking for 11 is a great way to see if there are 10 or more Jun 08 10:51:28 * Zakim sees ericP on the speaker queue Jun 08 10:51:30 Steve: can affect latency Jun 08 10:51:49 seems to me that this is extraordinarily common Jun 08 10:52:37 AndyS: can be either a query or a protocol issue Jun 08 10:54:03 q+ to say "but then we need a way to specify the snapshot of the database" Jun 08 10:54:03 * Zakim sees ericP, Yoshio on the speaker queue Jun 08 10:54:13 ack ericp Jun 08 10:54:13 ericP, you wanted to say that the SQL world was happy to hear "10 answers, don't know if there are more" and to say that asking for 11 is a great way to see if there are 10 or more Jun 08 10:54:16 * Zakim sees Yoshio on the speaker queue Jun 08 10:54:19 q+ do we want a percise upper bound or approx.? Jun 08 10:54:19 * Zakim KevinW, you typed too many words without commas; I suspect you forgot to start with 'to ...' Jun 08 10:54:19 EricP: arguing from the point of view of a straightforward implementation (both client and server) Jun 08 10:54:36 q+ Jun 08 10:54:36 * Zakim sees Yoshio, rob_ on the speaker queue Jun 08 10:54:43 ack yoshio Jun 08 10:54:43 Yoshio, you wanted to say "but then we need a way to specify the snapshot of the database" Jun 08 10:54:45 * Zakim sees rob_ on the speaker queue Jun 08 10:55:10 * DanC hears too much noise to understand yoshio Jun 08 10:55:13 (can't hear/understand_ Jun 08 10:55:35 yoshio, your voice is distorted Jun 08 10:55:44 Yoshio: but then we need a way to specify the snapshot of the database Jun 08 10:55:49 q+ Jun 08 10:55:49 * Zakim sees rob_, KevinW on the speaker queue Jun 08 10:56:03 ack rob Jun 08 10:56:03 * Zakim sees KevinW on the speaker queue Jun 08 10:56:11 I would not like the req to decide the (flag/number left) issue. There is little point to the req if it is foced to do the count as well. Jun 08 10:56:17 if the DB changes, the "11th" doesn't have any meaning Jun 08 10:56:35 RobS: what is understood a s a "result" Jun 08 10:56:40 +1 AndyS Jun 08 10:56:52 ack kevin Jun 08 10:56:52 * Zakim sees no one on the speaker queue Jun 08 10:57:16 DanC: ... matter of fact Jun 08 10:57:34 kevin: do we want an exact upper bound? could cause more impl burden on the server. Jun 08 10:58:11 DanC: does "upper bound" appeal to somebody Jun 08 10:58:25 to Kevin Jun 08 10:58:38 +1 on upper bound Jun 08 10:58:57 I like 3.10 as is Jun 08 10:59:07 * kendall too, fwiw. Jun 08 10:59:09 patH: DQL could, by requst, give a count of results Jun 08 10:59:33 I like 3.10 enough to support it. Jun 08 10:59:36 prefer 3.10a Jun 08 11:00:02 Pat - is that the SQL count() allowing aprox answers? If so, I see it as different Jun 08 11:00:03 I see 3.10 and 3.10a as disjoint Jun 08 11:00:03 +1 on 3.10a pref Jun 08 11:00:20 I agree with DaveB Jun 08 11:00:24 DanC: motivation to spend time on this is in discussing/ bounderaries with other groups Jun 08 11:00:29 AndyS: yes, that was it. Jun 08 11:00:32 yes, strictly speaking, 3.10 and 3.10a are not really variants. I think I'll separate them. Jun 08 11:00:34 3.10 & a they even have differnet titles :) Jun 08 11:00:45 * DaveB has to leave telcon Jun 08 11:01:07 yeah :) Jun 08 11:01:11 EricP's action continued (which one Eric?) Jun 08 11:01:11 +1 on split of 3.10 and 3.10a. I can see that they are disjoint. Jun 08 11:01:19 bye guys... Jun 08 11:01:21 There's probably some relationship between 3.10, 3.10a and streamability, I agree. Jun 08 11:01:25 -ChrisR Jun 08 11:01:45 <-- DaveB has quit (Quit: Client exiting) Jun 08 11:01:46 * alberto bye DaveB Jun 08 11:01:55 ==== ADJOURNED Jun 08 11:01:59 -DanC Jun 08 11:02:04 -PatH Jun 08 11:02:06 -Josd+DirkC Jun 08 11:02:06 EricP's action to provide a fwf registration is continued Jun 08 11:02:09 -Yoshio Jun 08 11:02:11 Any pointers on scribe protocol? Send links if any. Thanks. Jun 08 11:02:30 -HowardK Jun 08 11:03:43 Bye folks. Jun 08 11:03:51 <-- FarrukhNajmi has quit (Quit: Download Gaim: http://gaim.sourceforge.net/) Jun 08 11:03:56 * alberto leaving bye Jun 08 11:04:01 Cheers Jun 08 11:04:03 -Farrukh Jun 08 11:04:04 -??P15 Jun 08 11:04:06 pat leaving Bye Jun 08 11:04:09 <-- Yoshio (fuku@18.29.0.64) has left #dawg Jun 08 11:04:11 <-- phayes has quit (Quit: Client exiting) Jun 08 11:04:14 * DanC should make a link to scribe how-to from WG homepage Jun 08 11:04:17 Zakim, help Jun 08 11:04:17 Please refer to http://www.w3.org/2001/12/zakim-irc-bot for more detailed help. Jun 08 11:04:19 Some of the commands I know are: Jun 08 11:04:20 xxx is yyy - establish yyy as the name of unknown party xxx Jun 08 11:04:22 if yyy is 'me' or 'I', your nick is substituted Jun 08 11:04:24 xxx may be yyy - establish yyy as possibly the name of unknown party xxx Jun 08 11:04:26 I am xxx - establish your nick as the name of unknown party xxx Jun 08 11:04:28 xxx holds yyy [, zzz ...] - establish xxx as a group name and yyy, etc. as participants within that group Jun 08 11:04:30 xxx also holds yyy - add yyy to the list of participants in group xxx Jun 08 11:04:32 who's here? - lists the participants on the phone Jun 08 11:04:34 who's muted? - lists the participants who are muted Jun 08 11:04:36 mute xxx - mutes party xxx (like pressing 61#) Jun 08 11:04:38 unmute xxx - reverses the effect of "mute" and of 61# Jun 08 11:04:40 <-- alberto (alberto@81.74.1.226) has left #dawg Jun 08 11:04:40 is xxx here? - reports whether a party named like xxx is present Jun 08 11:04:42 list conferences - reports the active conferences Jun 08 11:04:43 this is xxx - associates this channel with conference xxx Jun 08 11:04:44 excuse us - disconnects from the irc channel Jun 08 11:04:45 I last learned something new on Date:2004/02/1701:04:28Date: 2004/02/17 01:04:28 Date:2004/02/1701:04:28 Jun 08 11:05:14 <-- howardk has quit (Quit: howardk) Jun 08 11:06:22 zakim, who's on the phone? Jun 08 11:06:22 On the phone I see Kevin, ??P8, ??P10, ??P12, RobS, EricP, Kendall_Clark Jun 08 11:06:35 rob, coudl you try to access http://www.w3.org/2002/09/wbs/35463/DAWGf2f2/ for me? Jun 08 11:07:16 looks like I can get through Jun 08 11:07:25 (registration form) Jun 08 11:07:39 aha, perhaps i have completed my action item Jun 08 11:07:49 -Kevin Jun 08 11:07:57 <-- KevinW has quit (Quit: User pushed the X - because it's Xtra, baby) Jun 08 11:08:29 patH couldn't get to it, but i heard him mutter something about password probs as he left. Jun 08 11:08:45 I needed a password to see the page Jun 08 11:08:59 that;s good. it's limited ot the WG Jun 08 11:09:09 kendall, is it possible to put something in there to help people with ride-sharing to/from airports? Jun 08 11:09:19 (sorry, not kendall, eric) Jun 08 11:13:47 oh, like "what airport are you flying to/from, and when?" yes, ericp, that would be cool Jun 08 11:13:49 got to go... Jun 08 11:13:54 hasta, rob_ Jun 08 11:13:56 See you Jun 08 11:14:00 -RobS Jun 08 11:15:22 i may as well add the flight/ride stuff -- i'll have this action item until the day after the f2f anywyas

Received on Wednesday, 16 June 2004 13:02:52 UTC