Chatlog 2011-03-08

From SPARQL Working Group
Jump to: navigation, search

See original RRSAgent log and preview nicely formatted version.

Please justify/explain all edits to this page, in your "edit summary" text.

14:54:44 <RRSAgent> RRSAgent has joined #sparql
14:54:44 <RRSAgent> logging to
14:54:46 <trackbot> RRSAgent, make logs world
14:54:46 <Zakim> Zakim has joined #sparql
14:54:48 <trackbot> Zakim, this will be 77277
14:54:48 <Zakim> ok, trackbot; I see SW_(SPARQL)10:00AM scheduled to start in 6 minutes
14:54:49 <trackbot> Meeting: SPARQL Working Group Teleconference
14:54:49 <trackbot> Date: 08 March 2011
14:54:51 <LeeF> zakim, this will be SPARQL
14:54:51 <Zakim> ok, LeeF; I see SW_(SPARQL)10:00AM scheduled to start in 6 minutes
14:54:53 <LeeF> Chair: LeeF
14:55:02 <LeeF> Agenda:
14:55:13 <Zakim> SW_(SPARQL)10:00AM has now started
14:55:17 <LeeF> Scribenick: pgearon
14:55:20 <Zakim> +??P7
14:55:25 <NickH> NickH has joined #sparql
14:55:54 <SteveH_> SteveH_ has joined #sparql
14:56:54 <NickH> Zakim, ??P7 is me
14:56:54 <Zakim> +NickH; got it
14:57:08 <NicoM> NicoM has joined #sparql
14:57:18 <Zakim> +LeeF
14:57:29 <Zakim> +[IPcaller]
14:57:36 <AndyS> zakim, IPCaller is me
14:57:36 <Zakim> +AndyS; got it
14:57:42 <Zakim> +NicoM
14:57:52 <Zakim> +kasei
14:58:43 <kasei> Zakim, mute me
14:58:43 <Zakim> kasei should now be muted
14:59:49 <Zakim> +[IPcaller]
14:59:52 <chimezie> chimezie has joined #sparql
14:59:57 <SteveH> Zakim, [IPcaller] is me
14:59:57 <Zakim> +SteveH; got it
15:00:10 <Zakim> +Sandro
15:00:20 <Zakim> +pgearon
15:00:31 <MattPerry> MattPerry has joined #sparql
15:00:48 <Zakim> + +1.216.368.aaaa
15:00:54 <chimezie> Zakim, +1.216.368.aaaa is me
15:00:54 <Zakim> +chimezie; got it
15:01:13 <Zakim> +MattPerry
15:01:43 <LeeF> zakim, who's on the phone?
15:01:43 <Zakim> On the phone I see NickH, LeeF, AndyS, NicoM, kasei (muted), SteveH, Sandro, pgearon, chimezie, MattPerry
15:02:25 <LeeF> topic: Admin
15:02:26 <LeeF> PROPOSED: Approve minutes at
15:03:02 <LeeF> RESOLVED: Approve minutes at
15:03:13 <LeeF> Next regular meeting: 2011-03-15 @ 14:00 UK / 10:00 EST (scribe: Andy S) NOTE DIFFERENT TIME FOR NON-US LOCATIONS
15:05:14 <bglimm> bglimm has joined #sparql
15:05:31 <LeeF> topic: To Last Call
15:06:33 <LeeF> bglimm, we are wondering if you intend to look at the formal section of SPARQL Query when you have a chance?
15:07:04 <pgearon> SteveH: Query in the same state as last time
15:07:12 <Zakim> +bglimm
15:07:25 <chimezie> Zakim, mute me
15:07:25 <Zakim> chimezie should now be muted
15:08:11 <pgearon> bglimm: will still have a look at the formal section soon
15:08:56 <pgearon> Not much happened. The main thing still outstanding is the BNode issue
15:09:11 <pgearon> If there's anything else that needs doing, then someone should point it out
15:09:32 <pgearon> I was under the impression that all outstanding issues were handled (except test data)
15:09:37 <kasei> right
15:10:18 <pgearon> for AndyS... That was a few weeks ago now, so my memory isn't clear, but I think so yes
15:11:04 <pgearon> LeeF: AndyS and kasei to be asked to review Update at their convenience
15:11:17 <bglimm> Zakim, mute me
15:11:17 <Zakim> bglimm should now be muted
15:11:58 <LeeF>
15:12:42 <pgearon> LeeF: some progress for "Protocol"
15:12:54 <kasei> Zakim, unmute me
15:12:54 <Zakim> kasei should no longer be muted
15:13:10 <pgearon> topic: Service Description
15:13:55 <LeeF> "Add link to discussion on empty graphs in the Update document (when such a section exists) "
15:14:47 <pgearon> for LeeF: no, I wasn't
15:15:34 <pgearon> kasei: service description describes dropping empty graphs
15:16:00 <pgearon> kasei: supporting empty graphs
15:16:11 <pgearon> LeeF: this needs to have some reference in the Update document
15:16:36 <LeeF>
15:17:21 <pgearon> kasei: Update doc describes dropping graphs. Not clear from this description if it describes supporting empty graphs
15:17:58 <pgearon> LeeF: Update may not need to say what is supported in a store. Service description can describe this
15:18:59 <LeeF>
15:19:14 <LeeF> ACTION: Lee to propose text for SD document explaining 3.4.7 sd:EmptyGraphs
15:19:15 <trackbot> Created ACTION-406 - Propose text for SD document explaining 3.4.7 sd:EmptyGraphs [on Lee Feigenbaum - due 2011-03-15].
15:19:43 <pgearon> In the DROP section I was trying to describe how a store would behave in the 2 situations: 1. supporting empty graphs. 2. Not recording empty graphs. I tried hard not to commit to saying what should/should not be supported
15:20:09 <LeeF> pgearon, thanks, I'll take a look at that too
15:20:28 <pgearon> kasei: would like someone to glance at conformance section for service description
15:21:16 <pgearon> kasei: conformance section says, "must include one triple". AndyS would like this to say, "must include at least one triple"
15:21:23 <LeeF> ACTION: Greg to change conformance section in SD to refer to including "at least" one triple...
15:21:23 <trackbot> Created ACTION-407 - Change conformance section in SD to refer to including "at least" one triple... [on Gregory Williams - due 2011-03-15].
15:21:24 <SteveH> +1 to that
15:21:33 <LeeF> ACTION: Andy to look over current conformance section in service description modulo "at least" change from ACTION-407
15:21:33 <trackbot> Created ACTION-408 - Look over current conformance section in service description modulo "at least" change from ACTION-407 [on Andy Seaborne - due 2011-03-15].
15:21:56 <pgearon> topic: HTTP protocol document
15:21:59 <kasei> Zakim, mute me
15:21:59 <Zakim> kasei should now be muted
15:22:11 <chimezie> Zakim, unmute me
15:22:11 <Zakim> chimezie should no longer be muted
15:22:57 <pgearon> LeeF: comments on relationship to service description. Also comment from Leigh Dodds with other operations that might be included
15:23:29 <pgearon> chimezie: Thinks that operations in POST are beyond scope, but we should have this discussion
15:23:47 <pgearon> LeeF: will come back to new operations
15:24:05 <pgearon> LeeF: SteveH still has pending review on this doc
15:24:18 <pgearon> LeeF: Entailment document
15:24:34 <bglimm> Zakim, unmute me 
15:24:34 <Zakim> bglimm should no longer be muted
15:24:53 <pgearon> LeeF: received review from Jay. Should mark this off on Wiki page. Still waiting on review from Clark and Parsia
15:25:29 <pgearon> LeeF: to bglimm, should we discuss D-entailment today?
15:26:29 <pgearon> LeeF: JSON doc waiting for reviewers to have some time
15:26:33 <bglimm> I reminded C&P (talked o Hector) and he'll try to push Evren who is doing the review
15:26:36 <chimezie> Zakim, mute me
15:26:36 <Zakim> chimezie should now be muted
15:27:40 <pgearon> LeeF: can people please consider tests as we get closer to LC
15:28:17 <pgearon> LeeF: also need to review existing tests, as these have taken a back seat to editorial issues recently
15:28:19 <AndyS> 405 is done
15:28:34 <bglimm> I think I had one, which I completed, but can't remember what it was
15:28:42 <AndyS> .. that was 405.
15:28:50 <bglimm> I'll search for it
15:28:56 <LeeF> trackbot, close ACTION-405
15:28:57 <trackbot> ACTION-405 Draft xml for revised Extending BGP matching section for query closed
15:28:58 <kasei> LeeF, I've completed 396 and 403
15:28:59 <bglimm> Thans
15:29:07 <bglimm> s/Thans/Thanks/
15:29:14 <LeeF> trackbot, close ACTION-396
15:29:14 <trackbot> ACTION-396 Contact chime on Dataset Update Protocol issue closed
15:29:28 <LeeF> trackbot, close ACTION-403
15:29:28 <trackbot> ACTION-403 Answer to with counterexample to JB-2 closed
15:29:37 <bglimm> 398 I also did
15:29:47 <bglimm> prepare test cases for approval
15:29:51 <LeeF> trackbot, close ACTION-398
15:29:51 <trackbot> ACTION-398 Prepare entailment tests for next week for approval closed
15:29:57 <bglimm> Axel looked at them too
15:30:06 <LeeF> topic: Blank nodes in DELETE templates
15:30:08 <pgearon> LeeF: Next topic. Blank nodes in DELETE templates
15:31:20 <pgearon> LeeF: resolution was for blank nodes to represent wildcards when deleting
15:31:39 <LeeF>
15:31:44 <pgearon> LeeF: recently SteveH, AndyS, pgearon found issues around implementation
15:32:13 <SteveH> andy didn't capture my preference which was make it an error
15:32:14 <pgearon> LeeF: AndyS also did work on how to delete lists without this "shortcut"
15:32:31 <LeeF> PROPOSED: SPARQL 1.1 Update forbids blank nodes in DELETE templates 
15:33:07 <SteveH> seconded
15:33:10 <bglimm> seconded
15:33:11 <pgearon> LeeF: does anyone have anything further to discuss on this, or serious concerns?
15:33:26 <LeeF> RESOLVED: SPARQL 1.1 Update forbids blank nodes in DELETE templates
15:33:43 <pgearon> Sure
15:34:00 <pgearon> I'll change it either way, but an action helps track it
15:34:02 <LeeF> ACTION: Paul to edit Update document to note that blank nodes in DELETE templates are an error
15:34:03 <trackbot> Created ACTION-409 - Edit Update document to note that blank nodes in DELETE templates are an error [on Paul Gearon - due 2011-03-15].
15:34:20 <LeeF> topic: Dataset protocol & service description 
15:35:14 <pgearon> LeeF: service description initially conceived as description of features and capabilities of endpoints supporting SPARQL protocol
15:35:45 <pgearon> LeeF: scope expanded to description of things like endpoint URLs
15:36:45 <pgearon> LeeF: do we need additions to service description to include dataset protocol. Preference is to not make large changes
15:36:49 <chimezie> Zakim, unmute me
15:36:49 <Zakim> chimezie should no longer be muted
15:37:15 <chimezie>
15:37:19 <chimezie>
15:37:23 <pgearon> chimezie: main comments are from Leigh Dodds and Rob Vesse
15:37:38 <pgearon> chimezie: Leigh wants to do GET/PUT on protocol
15:37:47 <LeeF> s/protocol/graph store
15:38:30 <LeeF> Leigh asks: how do you find the URL to POST new graphs to?
15:38:47 <pgearon> chimezie: Leigh has question on graph store being a dataset, but don't think it is
15:39:05 <LeeF> Leigh would also want to implement the dataset protocol without exposing SPARQL query or SPARQL update endpoints, but SD document assumes that you're talking about SPARQL protocol & not dataset protocol
15:39:20 <LeeF> Rob V says similar comments
15:40:01 <LeeF> Greg proposed including a RESTDataset class, but not more substantive changes at this time
15:40:06 <Zakim> -NickH
15:40:26 <LeeF> chimezie: there's currently an informative section in dataset protocol document talking about relationship with SD
15:40:33 <chimezie> sd:RESTDataset
15:40:33 <LeeF> ... could introduce vocabulary there
15:40:39 <chimezie> sd:defaultGraphStoreDescription
15:41:20 <kasei> I don't understand what sd:defaultGraphStoreDescription would mean.
15:41:26 <LeeF>
15:41:39 <Zakim> +??P7
15:41:46 <NickH> Zakim, ??P7 is me
15:41:46 <Zakim> +NickH; got it
15:42:05 <LeeF> q+ to ask if it really needs to support update
15:42:44 <pgearon> chimezie: current relationship between services and dataset in the service desc doc
15:42:52 <chimezie> sd:defaultDatasetDescription
15:43:07 <pgearon> chimezie: can't reuse this term because the range is Dataset
15:43:09 <LeeF> the range of that term is sd:Dataset
15:43:25 <pgearon> chimezie: my understanding is that a graph store is not a dataset
15:43:28 <AndyS> q+ to talk about g-boxes and g-snaps
15:43:28 <kasei> Zakim, unmute me
15:43:30 <Zakim> kasei should no longer be muted
15:43:38 <pgearon> chimezie: new term would be the same thing, but relates service to a graph store instead
15:43:51 <pgearon> LeeF: do we need to define a graph store class?
15:44:35 <LeeF> chimezie: the object of the sd:defaultGraphStoreDescription would be the URL against which dataset protocol operations should be sent
15:44:52 <LeeF> kasei: but it's the dataset protocol, not the graph store protocol?
15:45:14 <sandro> q+
15:46:01 <LeeF> what about sd:datasetProtocolLocation or something ?
15:46:02 <sandro> q+ to say I think a dataset can be mutable even if a graph isn't -- in changing it, you're changing what graphs are in it
15:46:15 <pgearon> chimezie: if we're going to say that a graph store is not a dataset then don't know if protocol or SD docs are the place to say that
15:46:25 <LeeF> ack AndyS
15:46:25 <Zakim> AndyS, you wanted to talk about g-boxes and g-snaps
15:47:23 <sandro> (but datasets in SPARQL *do* change, right?)
15:47:58 <kasei> isn't the dataset protocol about moving from one dataset to another?
15:48:05 <LeeF> sandro: don't all existing implementations of SPARQL 1.0 involve datasets changing?
15:48:41 <LeeF> AndyS: it's not the dataset which is changing, it's what the service is currently supporting that is changing
15:48:56 <LeeF> sandro: do people conceptualize this as replacing one dataset from another ?
15:49:06 <chimezie> thread on mutability of datasets and graphs:
15:49:15 <LeeF> AndyS: not necessarily, but in the same way that people using a programming language think of sets as mutable, even though mathematical sets are immutable
15:49:27 <LeeF> sandro: wonder if treating datasets as a mutable structure is a simple way to deal with this ?
15:49:35 <LeeF> AndyS: we call the mutable structure a graph store
15:50:07 <LeeF> sandro: seems somewhat confusing to draw this distinction
15:50:15 <LeeF> AndyS: changing update document to remove term "graph store" would be a big change
15:51:22 <pgearon> AndyS: the slight trick is that we don't put datasets on the web. We put graphs on the web
15:51:37 <pgearon> AndyS: don't have a theory of datasets as computational objects
15:51:40 <Zakim> -bglimm
15:52:21 <kasei> q+
15:52:25 <LeeF> AndyS: is it a dataset protocol? 
15:52:38 <Zakim> + +41.86.528.aabb
15:52:47 <bglimm> Zakim, +41.86.528.aabb is me
15:52:47 <Zakim> +bglimm; got it
15:52:58 <bglimm> Zakim, mute me
15:52:58 <Zakim> bglimm should now be muted
15:53:19 <LeeF> ack kasei
15:53:25 <AndyS> q-
15:53:44 <LeeF> kasei: hesitant to split up definition of SD across multiple documents
15:53:57 <LeeF> ... there's a lot of value to having it contained to one document
15:54:28 <LeeF> kasei: would sd:defaultGraphStoreDescription be used in addition to sd:defaultDatasetDescription for services that implement both protocols?
15:54:35 <LeeF> chimezie: the name probably isn't great
15:54:47 <LeeF> chimezie: my intuition is that it would use both
15:54:51 <sandro> q-
15:55:45 <LeeF> kasei: i thought we accomplish this already by typing an existing dataset as a RESTDataset
15:56:18 <LeeF> chimezie: we don't have an explicit axiom saying a RESTDataset is disjoint from a Dataset, but there's an implication that they are disjoint, so it would be confusing to use them in that way
15:56:37 <LeeF> kasei: we may run into problems because people are not going to understand that they're disjoint
15:56:53 <LeeF> kasei: i think of one as a mutable version of the other, or an access point to move from one dataset to the next
15:57:28 <LeeF> chimezie: via Leigh's comments, the question is out there as to the relationship between a graph store and a dataset
16:00:29 <LeeF> q-
16:02:03 <NickH> is it decided if a GET to a SPARQL endpoint returns the service description?
16:02:15 <kasei> NickH, yes
16:02:30 <Zakim> -chimezie
16:02:32 <MattPerry> bye
16:02:33 <Zakim> -LeeF
16:02:37 <Zakim> -NicoM
16:02:40 <Zakim> -bglimm
16:02:42 <Zakim> -AndyS
16:02:43 <Zakim> -kasei
16:02:43 <Zakim> -MattPerry
16:02:45 <Zakim> -pgearon
16:02:49 <NickH> kasei: ok, thanks
16:02:54 <Zakim> -NickH
16:02:59 <Zakim> -SteveH
16:03:09 <LeeF> LeeF: With chair hat on, I'm very tempted to say that given resource and experience constraints, that SD is only about SPARQL protocol and is unrelated to the dataset protocol
16:03:38 <LeeF> NickH, yes, see
16:03:55 <LeeF> pgearon, can you do the magic to generate the minutes, please?
16:04:03 <pgearon> sure
16:04:07 <LeeF> many thanks
16:06:44 <NickH> LeeF: thanks, just checking if that has been confirmed