14:28:08 <RRSAgent> RRSAgent has joined #dawg
14:28:08 <RRSAgent> logging to http://www.w3.org/2005/04/12-dawg-irc
14:28:17 <Zakim> +JeenB
14:28:21 <ericP> zakim, aaaa is Monte
14:28:21 <Zakim> +Monte; got it
14:28:29 <Zakim> +DanC
14:28:52 <ericP> zakim, Monte holds SteveH, DaveB, AlberertoR, EricP
14:28:52 <Zakim> +SteveH, DaveB, AlberertoR, EricP; got it
14:28:58 <DanC> DanC has changed the topic to: RDF Data Access http://www.w3.org/2001/sw/DataAccess/ 12 Apr
14:29:12 <Zakim> +??P21
14:29:31 <DanC> Zakim, Monte holds AlbertoR, DaveB, 
14:29:31 <Zakim> DaveB was already listed in Monte, DanC
14:29:32 <Zakim> +AlbertoR; got it
14:29:36 <HiroyukiS> Zakim, ??P21 is HiroyukiS
14:29:36 <Zakim> +HiroyukiS; got it
14:29:44 <Zakim> +??P22
14:29:48 <AndyS> zakim, ??P22 is AndyS
14:29:48 <Zakim> +AndyS; got it
14:29:51 <Zakim> +Kendall_Clark
14:29:59 <DanC> Zakim, who is on the phone?
14:29:59 <Zakim> On the phone I see Monte, JeenB, DanC, HiroyukiS, AndyS, Kendall_Clark
14:30:00 <Zakim> Monte has AlbertoR
14:30:10 <DanC> Zakim, Monte holds SteveH, DaveB, AlberertoR, EricP
14:30:10 <Zakim> +SteveH, DaveB, AlberertoR, EricP; got it
14:30:14 <DanC> Zakim, who is on the phone?
14:30:14 <Zakim> On the phone I see Monte, JeenB, DanC, HiroyukiS, AndyS, Kendall_Clark
14:30:15 <Zakim> Monte has SteveH, DaveB, AlberertoR, EricP
14:30:49 <DanC> Zakim, take up item 1
14:30:49 <Zakim> agendum 1. "Convene, take roll, review records and agenda http://lists.w3.org/Archives/Public/public-rdf-dawg/2005AprJun/0059.html" taken up [from DanC]
14:30:54 <Zakim> +Kevin
14:31:09 <DanC> Regrets: BryanT
14:31:19 <DanC> Regrets: BryanT, Howard
14:31:45 <DanC> Zakim, who is on the phone?
14:31:45 <Zakim> On the phone I see Monte, JeenB, DanC, HiroyukiS, AndyS, Kendall_Clark, Kevin
14:31:47 <Zakim> Monte has SteveH, DaveB, AlberertoR, EricP
14:32:00 <DanC> Scribe: DanC
14:32:17 <DanC> -> http://www.w3.org/2005/04/05-dawg-minutes.html 5 Apr minutes
14:32:30 <DanC> RESOLVED to approve 5 Apr minutes
14:32:41 <DanC> next meeting: 19 Apr. scribe volunteer?
14:32:52 <DanC> Zakim, pick a scribe
14:32:52 <Zakim> Not knowing who is chairing or who scribed recently, I propose DanC
14:33:04 <DanC> next meeting: 19 Apr. Jeen to scribe
14:33:31 <DanC> (4 actions from the agenda continued)
14:33:55 <DanC> agenda + graphSolutionMapping 
14:33:59 <DanC> Zakim, next item
14:33:59 <Zakim> agendum 2. "issue: fromUnionQuery" taken up [from DanC]
14:34:31 <kendall> No decisions made to close yet; been moved from QL to protocol.
14:34:48 <kendall> Getting comments like 'why are from/with gone'?
14:35:36 <DaveB> DaveB has joined #dawg
14:35:46 <DanC> KC: protocol editor's draft refers to DataSet and describes it in WSDL...
14:35:51 <DanC> ... but there aren't HTTP examples yet
14:36:29 <DaveB> I can scribe now if you like...
14:37:01 <kendall> I don't understand the comment about "expected something about background graph"
14:37:10 <DanC> -> http://lists.w3.org/Archives/Public/public-rdf-dawg/2005AprJun/0054.html Re: fromUnionQuery
14:37:20 <Zakim> +Pat_Hayes
14:37:32 <AndyS> Protocol doc says "The RDF dataset is a sequence of any number of URIs, each of which identifies an RDF graph."
14:37:47 <db-scribe> discussion of background graph in protocol
14:37:58 <patH> patH has joined #dawg
14:38:14 <kendall> Right,  that's what I understood, Andy.
14:38:23 <AndyS> No distinguished background graph?
14:38:47 <db-scribe> there used to be two different keywords LOAD and FROM
14:39:07 <db-scribe> for this conv, FROM for background and FROM NAMED for named graph
14:39:18 <db-scribe> for this conversation, let's use^ (AndyS)
14:39:46 <db-scribe> KC -had tried to reflect the design into protoocl
14:39:55 <SteveH> q+ to propose a design without a distinguished background graph
14:40:04 <DanC> ack steveh
14:40:04 <Zakim> SteveH, you wanted to propose a design without a distinguished background graph
14:40:48 <AndyS> (can't hear Steve - keeps dropping out)
14:41:20 <db-scribe> maybe http://lists.w3.org/Archives/Public/public-rdf-dawg/2005JanMar/0440.html
14:41:23 <AndyS> http://lists.w3.org/Archives/Public/public-rdf-dawg/2005JanMar/0440.html
14:42:37 <db-scribe> SH - idea is instead of a privelged background graph,
14:42:48 <db-scribe> specify a set of graphs that are used if you don't use the bg grap
14:42:51 <db-scribe> h
14:43:24 <db-scribe> when not using the GRAPH keyword, it allows results from the lists of defaulted graphs
14:43:38 <db-scribe> rather than just from the bg graph
14:43:39 <kendall> wow, bad comms today
14:43:42 <AndyS> Is this like having trusted/untrusted flag on each graph?
14:44:02 <kendall> ah, they're in tuscany. the bastards. :>
14:44:41 <db-scribe> 'use' and 'constrain' - use is the list of graphs to use without the GRAPH keyword
14:44:45 <db-scribe> and constrain if you do
14:45:00 <db-scribe> there does not have to be a background graph
14:45:10 <db-scribe> the used grraphs are not the bg graph, they are still named
14:45:23 <kendall> +1 for explicit better than implicit
14:45:46 <db-scribe> SH - there are more than 1, as many as you like, explicit URIs for non-GRAPH triple matches
14:46:43 <db-scribe> sh - background graph has no name
14:46:49 <db-scribe> as well as being a single graph
14:46:50 <AndyS> Seems to restrict datasets to where BG graph is also named.
14:47:07 <db-scribe> whereas the use list of graphs allows a set of graph to be consulted
14:47:33 <AndyS> i.e. BG graph is one of the named graphs - all union/merge issues is separable issue
14:47:38 <AndyS> ??
14:47:52 <db-scribe> there are 2 sets of graphs - (set of use uris)  and (set of consult uris) graphs
14:48:06 <db-scribe> the former are used without GRAPH, the latter with GRAPH name
14:48:14 <AndyS> Test case?
14:48:41 <kendall> fwiw, the current protocol working draft doesn't say anything about a background graph because I understand "background graph" in the QL spec to be implicit, unnamed, and in fact unnameable by a client/requester...
14:48:58 <db-scribe> union of (set of use graphs) is the bg graph?
14:49:30 <db-scribe> the (set of use graphs) are not merged
14:51:08 <db-scribe> DC - (set of use graphs) merged are the bg graph and all the (use, consult) graphs are named graphs and can be refered to
14:51:13 <kendall> +1 andy
14:51:44 <db-scribe> SH - having problems with the bg graph having no name in a use case
14:52:03 <kendall> zakim, mute me
14:52:03 <Zakim> Kendall_Clark should now be muted
14:52:18 <db-scribe> SH - case is data management of the background graph
14:52:37 <db-scribe> dataset covers things the protocol can't specify
14:52:43 <db-scribe> ^DC-
14:52:50 <kendall> zakim, unmute me
14:52:50 <Zakim> Kendall_Clark should no longer be muted
14:54:11 <db-scribe> AS - hearing most it being impl issues
14:54:40 <kendall> I'm interested too
14:54:57 <db-scribe> SH - this is a new use case
14:55:14 <DanC> ACTION SteveH: write up use case that can't be handled by current design (?)
14:55:27 <db-scribe> ^- data management of background graph
14:56:15 <db-scribe> DC - what keyword names...
14:56:35 <db-scribe> KC - bg graph implicit
14:57:06 <db-scribe> 2 sequences of uris that need recording
14:57:13 <db-scribe> KC to think about names of params
14:57:37 <db-scribe> DC some apps allow you to load the bg graph, some don't
14:58:15 <AndyS> Dataset is "{ G, (u1, G1), (u2, G2), . . . (un, Gn) }" NB the first G has no URI.
14:58:48 <kendall> (I called it the "target graph" instead of background, because that doesn't imply, to me at least, implicitness)
14:59:09 <db-scribe> ask about where does the data go - all data into bg graph (JanneS)
14:59:22 <db-scribe> for rasqal not quite impl (DaveB)
14:59:43 <db-scribe> for arq can do either (AndyS) via the query language not protocol
15:00:25 <db-scribe> how about 'data' and 'background'
15:00:50 <db-scribe> DC - could call them trusted and untrusted
15:00:54 <db-scribe> AS - have tried to avoid that
15:01:04 <kendall> "background" & "foreground"
15:01:09 <kendall> "named" & "unnamed"
15:01:14 <kendall> "named" & "default"
15:01:32 <db-scribe> DaveB: I wouldn't use naming for fore/backgroun distinction
15:02:01 <AndyS> +1 to "named" and "default"
15:03:24 <db-scribe> KC - can have this in both query and protoocl, need to have a way to handle when they don't agree
15:04:11 <db-scribe> hmm, maybe I prefer "data" and "default" as param names
15:04:32 <db-scribe> KC - 0 or more URIs for 2 lists - named and default
15:05:15 <db-scribe> DC requests http example
15:05:24 <db-scribe> of the design KC's working on
15:05:48 <DanC> ACTION KC: draft protocol section on specifying datasets (fromUnionQuery) incl WSDL types and HTTP examples
15:06:08 <db-scribe> EricP offers to impl
15:06:43 <DanC> ACTION EP: try out "2 list of URIs" data sets design.
15:08:36 <db-scribe> q+
15:08:42 <DanC> ack db-scribe 
15:10:10 <db-scribe> about getting this in the QL also
15:13:14 <AndyS> How about: (does not scale)   #!/bin/sparql --default=URI1 --named=URI2 --named=URI3
15:13:36 <db-scribe> passing along (data to query+query) together in a document is very handy - xquery and rdf query
15:14:10 <db-scribe> DC - analogy with xslt also
15:14:24 <db-scribe> DC (earlier) said this was discussed at f2f5 boston
15:14:40 <DanC> (no, analogy with xslt does seem new)
15:15:06 <db-scribe> KC - rdf dataset in protocol and/or query language to be pretyt compelling
15:15:16 <db-scribe> .. need to say what happens when they disagree
15:15:22 <SteveH> I dont care where it is specified. but I onyl want it in one place
15:15:42 <db-scribe> kevinW - would b enice to have it in tke queyr lang
15:15:52 <db-scribe> .. notion of subqueries, kinda harder with a protocol
15:16:10 <db-scribe> a forward compat. point, e.g. naming result of a describe
15:16:34 <db-scribe> SH - why would this be harder in the protocol?
15:16:37 <DanC> (chairing thought: making decisions that don't completely close an issue is hardly worth the bother)
15:17:10 <AndyS> For DAWG2 ...  FROM CONSTRUCT  or  FROM DESCRIBE c.f. SQL FROM (SELECT ...)
15:17:13 <db-scribe> KW - expressing a query as the result of another query
15:18:07 <db-scribe> KC - would expect comments on this change
15:18:51 <kendall> DC -- It seems "invisible" and "surprising" because it's not clear to people how to express the RDF dataset in the protocol docs. 
15:19:05 <db-scribe> SH - unahappy that the dataset appear in two places, want it in exactly 1
15:19:41 <kendall> KC -- I don't agree with this, because the comments people have made about surprise & visibility have been comments on the query language spec. To wit, "I liked the QL better when it contained FROM/WITH."
15:20:39 <AlbertoR> I disagree - generate an error if in both might be not good - better to allow them in both and overrride
15:20:55 <kendall> well, they're standing next to each other, so we may not need to hear it all :>
15:21:19 <kendall> FWIW, I'm -1, I think, on it being an error for the rdf-dataset to "be in both". If they agree, who cares?
15:21:37 <db-scribe> SH and DB discussed it - haiving it in either is ok, having in both is an error.
15:21:47 <db-scribe> would work, SH slightly  uncomforatable
15:21:53 <kendall> +1 Alberto's point
15:22:04 <db-scribe> Alberto - allow both, let protocol override
15:22:11 <DanC> it appeals to me to have the protocol override the ql
15:22:16 <db-scribe> ... or the other way round
15:22:17 <db-scribe> !
15:23:12 <db-scribe> SH, EP - rather not have both -w ould confuse users if both ql+protocol had graphs specified
15:23:41 <db-scribe> KC - error if in both and they don't agree?
15:23:52 <db-scribe> EP - slighly scared by that
15:23:56 <AlbertoR> the problem to me is query.sparql might generate an error if run locally (API/non-protocol) - and OK if run through HTTP/Apache for example - the user would not get that
15:23:59 <db-scribe> ... multiple names for same resource
15:24:34 <db-scribe> canonicalisation of sources 
15:24:48 <db-scribe> order of lists of uris?
15:24:59 <db-scribe> EP - no reason for order to matter
15:25:03 <kendall> "same set of strings"
15:25:05 <DanC> options: * only in protocol * only in QL * in either, error if both * in either, QL overrides protocol * in either, protocol overrides QL
15:25:22 <DanC> Zakim, who is on the phone?
15:25:22 <Zakim> On the phone I see Monte, JeenB, DanC, HiroyukiS, AndyS, Kendall_Clark, Kevin, Pat_Hayes
15:25:24 <Zakim> Monte has SteveH, DaveB, AlberertoR, EricP
15:25:55 <DanC> options: * only in protocol * only in QL * in either, error if both * in either, QL overrides protocol * in either, protocol overrides QL * in either, error if they disagree
15:26:42 <kendall> dc: only in proto
15:26:45 <kendall> jeen: last
15:27:18 <DanC> HS: both
15:27:20 <kendall> hiroyuki: "in both"
15:27:32 <kendall> Andy: protocol overrides QL
15:27:48 <kendall> Kendall: last, same as Jeen
15:28:14 <kendall> Kevin: LEAVE ME ALONE!
15:28:30 <kendall> PatH: No opinion. Why are we discussing this?
15:28:37 <kendall> (because why would there be a clash?)
15:28:40 <DanC> Kevin: in either. don't care how conflicts are resolved
15:29:07 <db-scribe> SH - in either, error if used in both simultaneous
15:29:13 <kendall> SteveH: in either, error in both
15:29:23 <kendall> Dave: in either, don't care about conflict resolution
15:29:50 <kendall> Eric: No op.
15:31:04 <kendall> graham's point is good, if it stands
15:31:17 <kendall> (er, that was my opinion only)
15:31:35 <db-scribe> DaveB
15:31:42 <kendall> Kendall
15:31:54 <ericP> apathy rules!
15:33:32 <db-scribe> DC - how to pick what happens when the conflic?
15:33:44 <AndyS> Use case: have a production query which has a target in QL.  Now want to test on new (different) dataset.  Protocol overrides query.
15:33:50 <db-scribe> when the query and protocol datasets conflict
15:34:03 <DanC> ACTION AFS: add syntax to specify datasets in the QL
15:34:31 <DanC> Zakim, next item
15:34:31 <Zakim> agendum 3. "issue: valueTesting" taken up [from DanC]
15:34:55 <DanC> ACTION: DaveB to work with EricP to clarify valueTesting proposal ETA 19 Apr
15:35:05 <DanC> ACTION: HowardK to seek clarification on valueTesting w.r.t. differenced from XQuery
15:36:33 <db-scribe> discussion of GK comment on dates in valeuTesting
15:36:40 <kendall> zakim, mute me
15:36:40 <Zakim> Kendall_Clark should now be muted
15:36:51 <DanC> Zakim, next agendum
15:36:51 <Zakim> agendum 4. "issue: punctuationSyntax" taken up [from DanC]
15:36:59 <kendall> thx dave
15:37:19 <kendall> zakim, unmute me
15:37:19 <Zakim> Kendall_Clark should no longer be muted
15:38:07 <db-scribe> I was impl the rq23 and got lost in the optional ';' and '.'s  etc. near end of }
15:38:24 <db-scribe> DC - saw no trogn supprot for reifications horthand, last week
15:38:28 <db-scribe> .. no strong ...
15:39:19 <db-scribe> DaveB - .'s at the end of a qname as a special case is tricky
15:39:50 <DanC> sounds like Dave's done as much as he intended to do re his turtle investigation action
15:40:32 <DanC> ericp seems to argue for { ex:s ex:p ex:o.boy .} 
15:40:37 <DanC> ericp seems to argue for { ex:s ex:p ex:o.boy. .} 
15:40:45 <db-scribe> yeah, where needed to resolve ambiguity
15:40:48 <DanC> ericp seems to argue against { ex:s ex:p ex:o.boy.} 
15:40:53 <AndyS> q+ on chars in qnames (not dots)
15:40:54 <db-scribe> and *also* editing Turtle to alllow '.'s
15:43:15 <db-scribe> so I'm happy to align turtle with expanded qnames re xml11
15:43:37 <DanC> PROPOSED: to not have reification syntax
15:43:48 <db-scribe> seconded ericp
15:44:01 <db-scribe> objections: none
15:44:15 <db-scribe> abstain: KC, AS, PH, KW
15:44:19 <DanC> Zakim, who's on the phone?
15:44:19 <Zakim> On the phone I see Monte, JeenB, DanC, HiroyukiS, AndyS, Kendall_Clark, Kevin, Pat_Hayes
15:44:22 <Zakim> Monte has SteveH, DaveB, AlberertoR, EricP
15:44:41 <DanC> abstain: KC, AS, PH, KW, JB
15:45:25 <DanC> abstain: KC, AFS/KW, PH, JB
15:46:03 <DanC> in favor: Dan/w3c, HiroyukiS, SteveH, DaveB, AlbertoR
15:46:21 <DanC> so RESOLVED.
15:46:35 <DanC> ACTION AFS: take it out
15:47:03 <DanC> Zakim, next agendum
15:47:03 <Zakim> agendum 5. "issue sort" taken up [from DanC]
15:47:13 <DanC> . ACTION: PatH to review sort design
15:48:11 <db-scribe> PH - reviewed it, it's complicated, but it would have to be
15:48:15 <DanC> (action done to my satisfaction)
15:48:41 <db-scribe> PH - need a total ordering
15:48:54 <db-scribe> .. wondering about the motivation was
15:49:22 <db-scribe> AS - made a total ordering in response to earlier PH feedback, but might have gone too far (?)
15:49:35 <db-scribe> PH - could be simplified if a TO's not needed
15:50:34 <db-scribe> noted that sql doesn't have this req, but requires a consistent ordering 
15:50:38 <db-scribe> in impls
15:51:27 <AndyS> Suggest choose simpler case based on "Case 1: arbitrary consistent ordering"
15:51:27 <db-scribe> EP - impl dep oreding only affects "getting the next 10 results" and if you switched an impl between then
15:51:50 <db-scribe> DC- call for test case writer for ordering
15:52:23 <db-scribe> DC about 10 cases - 6 normal cases, 4 weird things users want to do
15:53:18 <db-scribe> JB offers
15:53:21 <AndyS> Thanks Jeen
15:53:37 <db-scribe> JB askes about ordering in test case results
15:53:54 <db-scribe> AS - suggest where we case, we do the results in XML
15:54:04 <db-scribe> which preserves the order
15:54:31 <DanC> (aha... I didn't realize that the sort issue impacted the xml results set format spec)
15:54:44 <db-scribe> DB - not sure if vbr xml format says orderred or not, probably follows query model.
15:55:01 <kendall> this telcon has had a bit of everything!
15:55:13 <db-scribe> (5 yr old interrupts play)
15:55:31 <db-scribe> SH - if we added sorting, was proposing to add an index to the testcases of the order of the result
15:57:46 <db-scribe> SH - would be irrelvant if there is no ORDER BY
15:57:46 <db-scribe> ,.. correction, only there if there is an ORDER BY
15:57:48 <AndyS> I have a corner case ... :-)
15:57:49 <db-scribe> AS -worry about doing subgraph isomorphism vs graph isomorphism in comparing results.  (scribe?)
15:57:51 <db-scribe> SH - can 't number them if you don't have order by;  must number them if there is order by
15:57:55 <db-scribe> DC - would be a new indexProperty on the bnode with the order
15:58:09 <db-scribe> DC - doesn't sound enough energy for sorting in group
15:58:26 <db-scribe> SH - at present sounds like research, in favour not doing sorting
15:58:29 <DanC> ACTION JB: write a few sort tests
15:58:36 <DanC> Zakim, next agendum
15:58:36 <Zakim> agendum 6. "SPARQL QL publication candidate" taken up [from DanC]
15:59:28 <DanC> done: ACTION: DaveB to review rq23 editor's draft this week
15:59:37 <DanC> EP and AFS ack daveb's comments and bijan's
15:59:38 <db-scribe> EP, AS - not caught up yet
15:59:55 <db-scribe> EP planning to bug me till I explain my comments
16:00:01 <DanC> ACTION: PatH to internal review of rq23, starting monday
16:00:22 <DanC> done: ACTION: AndyS to clarify 5.4 w/r/t closed world assumption
16:02:26 <db-scribe> DC - 19Apr pub date realistic?
16:03:52 <ericP> "editorial in the loosest possible sense": changes will be done with an editor
16:04:11 <db-scribe> AS - only likely missing thing that's substantial is the fromUnionQuery
16:04:37 <DanC> PROPOSED: to publish editor's best effort this week
16:04:55 <db-scribe> (this is not a LC I assume)
16:04:55 <kendall> zakim, unmute me
16:04:55 <Zakim> Kendall_Clark was not muted, kendall
16:05:02 <DanC> PROPOSED: to publish editor's best effort this week, publication date early next week
16:05:48 <db-scribe> objections: none
16:05:52 <db-scribe> abstentions: SH
16:06:00 <db-scribe> resolved
16:06:00 <DanC> agenda?
16:06:07 <Zakim> -Kendall_Clark
16:06:08 <Zakim> -Kevin
16:06:10 <KevinW> KevinW has left #DAWG
16:06:18 <Zakim> -HiroyukiS
16:06:23 <DanC> jeen, are you in Tuscany?
16:06:28 <jeen> nope
16:06:31 <db-scribe> :)
16:06:54 <AlbertoR> AlbertoR has left #dawg
16:06:56 <DanC> ok... re scribe duties, there's a "scribe tips" link from the WG homepage to http://lists.w3.org/Archives/Public/public-rdf-dawg/2004AprJun/0624.html
16:07:17 <jeen> ok thx
16:07:36 <DaveB> oh, are you doiing it jeen?
16:07:48 <jeen> dave no next week
16:07:54 <DaveB> ah, fine
16:09:43 <Zakim> -JeenB
16:10:05 <jeen> jeen has left #dawg
16:11:02 <Zakim> -DanC
16:11:22 <Zakim> -Pat_Hayes
16:11:25 <Zakim> -Monte
16:11:58 <Zakim> -AndyS
16:11:59 <Zakim> SW_DAWG()10:30AM has ended
16:12:01 <Zakim> Attendees were +39.057.584.aaaa, JeenB, DanC, SteveH, DaveB, AlberertoR, EricP, AlbertoR, HiroyukiS, AndyS, Kendall_Clark, Kevin, Pat_Hayes
16:22:56 <DanC> Regrets+ Yoshio
16:23:06 <DanC> RRSAgent, make logs world-access
16:28:11 <DanC> s/ doesn't sound enough energy for sorting in group/ does sound enough energy for sorting in group
16:28:54 <DanC> I wanted to do a poll on postponing sort, but ran out of time
16:41:49 <patH> Yes, sorry, Dan, I maybe should have sounded less enthusiastic. In fact I am not very enthusiastic about sorting.
16:44:49 <DanC> I'm struggling to get this WG to close issues.
16:45:17 <DanC> we were down to 2 issues between us and a query language last call, as of the boston meeting. not only have we not closed those, we've re-opened others.
16:45:30 <DanC> sigh.
16:47:02 <DanC> oh well. we seem to be getting good comments
17:04:18 <AndyS> Significant user pull - implementers have to do much work.
17:05:09 <AndyS> "Top-10" use case is rather significant (but the chater says "select subgraph"!)
18:30:29 <Zakim> Zakim has left #dawg
18:32:37 <DanC> RRSAgent, bye
18:32:37 <RRSAgent> I see 9 open action items:
18:32:37 <RRSAgent> ACTION: SteveH to write up use case that can't be handled by current design (?) [1]
18:32:37 <RRSAgent>   recorded in http://www.w3.org/2005/04/12-dawg-irc#T14-55-14
18:32:37 <RRSAgent> ACTION: KC to draft protocol section on specifying datasets (fromUnionQuery) incl WSDL types and HTTP examples [2]
18:32:37 <RRSAgent>   recorded in http://www.w3.org/2005/04/12-dawg-irc#T15-05-48
18:32:37 <RRSAgent> ACTION: EP to try out "2 list of URIs" data sets design. [3]
18:32:37 <RRSAgent>   recorded in http://www.w3.org/2005/04/12-dawg-irc#T15-06-43
18:32:37 <RRSAgent> ACTION: AFS to add syntax to specify datasets in the QL [4]
18:32:37 <RRSAgent>   recorded in http://www.w3.org/2005/04/12-dawg-irc#T15-34-03
18:32:37 <RRSAgent> ACTION: DaveB to work with EricP to clarify valueTesting proposal ETA 19 Apr [5]
18:32:37 <RRSAgent>   recorded in http://www.w3.org/2005/04/12-dawg-irc#T15-34-55
18:32:37 <RRSAgent> ACTION: HowardK to seek clarification on valueTesting w.r.t. differenced from XQuery [6]
18:32:37 <RRSAgent>   recorded in http://www.w3.org/2005/04/12-dawg-irc#T15-35-05
18:32:37 <RRSAgent> ACTION: AFS to take it out [7]
18:32:37 <RRSAgent>   recorded in http://www.w3.org/2005/04/12-dawg-irc#T15-46-35
18:32:37 <RRSAgent> ACTION: JB to write a few sort tests [8]
18:32:37 <RRSAgent>   recorded in http://www.w3.org/2005/04/12-dawg-irc#T15-58-29
18:32:37 <RRSAgent> ACTION: PatH to internal review of rq23, starting monday [9]
18:32:37 <RRSAgent>   recorded in http://www.w3.org/2005/04/12-dawg-irc#T16-00-01