Chatlog 2009-04-28

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.

<LeeF> Present: Lee, Alex, Axel, Andy, iv_an_ru, IvanH, chimezie, john-l, lukewm, steveh, dnewman2, bijan, kjetil, prateek, kasei, Simon, pgearon, Janne
13:54:19 <trackbot> Meeting: SPARQL Working Group Teleconference
13:54:19 <trackbot>  Date: 28 April 2009
13:54:29 <LeeF> Chair: LeeF
13:54:34 <LeeF> Regrets: Orri
14:00:36 <LeeF> Scribe: chimezie
14:00:41 <LeeF> Scribenick: chimezie
14:04:47 <LeeF> topic: administrative
14:04:26 <chimezie> LeeF: we have quite a bit to cover today perhaps we can have a 90 minute call?
14:04:40 <LeeF> agenda -
14:04:54 <LeeF> PROPOSED: Approve minutes at
14:05:02 <SteveH> I had 1 issue with the mins
14:05:46 <SteveH> I think it was this: Eric Prud'hommeaux: it matters for portability between scripting and the bigger impls. [ Scribe Assist by Greg Williams ]
14:05:52 <chimezie> LeeF: perhaps we can update the minutes after the teleconference?
14:06:34 <chimezie> ... we can approve the minutes modulo this change
14:06:45 <LeeF> RESOLVED: Approve minutes at, modulo change noted by SteveH
14:06:54 <LeeF> ACTION: LeeF to talk to Eric to confirm minutes change from April 21
14:06:54 <trackbot> Created ACTION-10 - Talk to Eric to confirm minutes change from April 21 [on Lee Feigenbaum - due 2009-05-05].
14:08:07 <chimezie> LeeF: We will skip next weeks teleconference and resume a week from tommorow
14:08:27 <chimezie> Zakim, mute me
14:08:27 <Zakim> Chimezie_Ogbuji should now be muted
14:08:28 <LeeF> topic: introductions - Paul Gearon
14:09:12 <chimezie> Paul: I was one of the original developers on Mulgara.  Currently working with Fedora Commons with Mulgara
14:09:37 <chimezie> ... originally working on a storage system and query over the database.  Worked on first implementation of SPARQL
<LeeF> topic: Actions
14:10:27 <LeeF> open actions -
14:10:47 <chimezie> LeeF: 3 open actions having to do with rdf:text.  We will discuss this today (shortly)
14:11:05 <chimezie> ... Any additional review?
14:11:12 <LeeF> trackbot, close action-7
14:11:12 <trackbot> ACTION-7 Send a pointer to the mailinglist for rdf:text, when it's up to LC closed
14:11:14 <chimezie> SteveH: I consider it discharged
14:11:16 <LeeF> trackbot, close action-8
14:11:16 <trackbot> ACTION-8 Review rdf:text closed
14:11:18 <LeeF> trackbot, close action-9
14:11:18 <trackbot> ACTION-9 Try to review rdf:text closed
14:11:47 <LeeF> topic: rdf:text
14:12:42 <chimezie> LeeF: Jointly put forward by two WGs.  Potentially impacts SPARQL.  Andy has reviewed, so has SteveH.  As a group we need to decide how to respond
14:12:56 <AxelPolleres> q+ on status of the review.
14:13:14 <chimezie> LeeF: Best way forward on this?
14:13:24 <LeeF> ack AxelPolleres
14:13:24 <Zakim> AxelPolleres, you wanted to comment on status of the review.
14:14:02 <chimezie> AxelPolleres: Andy's comments are substantial and important cna can be addressed w/out to many changes.  rdf:text should not discuss semantic equivalent besides D-entailment.  
14:14:46 <Zakim> +dnewman2
14:14:55 <chimezie> ... We probably need a few mail cycles to finalized.  Perhaps a short agenda item during the F2F
14:15:25 <SteveH> +1 to having it in record
14:15:29 <chimezie> LeeF: Can we ensure the discusdsion is mentioned on the WG list for the benefit of everyone else?
14:15:45 <chimezie> AxelPolleres: Agreed.  
14:16:31 <bijan> I'm trying to review the comments
14:16:34 <bijan> I don't fully understand them
14:16:34 <dnewman2> dnewman2 has joined #sparql
14:16:35 <chimezie> LeeF: Does RIF/OWL need official responses? Anyone here have issues with Andy/SteveH speaking on behalf of the WG?
14:16:46 <bijan> q+
14:16:51 <chimezie> SteveH: Andy's understanding is slightly better than mine
14:16:53 <bijan> zakim, unmute me
14:16:53 <Zakim> bijan was not muted, bijan
14:17:23 <chimezie> AxelPolleres: We should try to clarify the last point regarding datatypes and ??string function??
14:17:49 <chimezie> bijan: We are either working up an official response or send a personal comment.  Please clarify
14:18:11 <SteveH> I feel that SPARQL should send an official response
14:18:12 <chimezie> LeeF: The WG hasn't sent an official repsonse, but we have a close relationship with these groups.  I'm happy with current process.
14:18:27 <chimezie> ... Not sure if anyone else wants a more formal process?
14:18:31 <iv_an_ru> IMHO a personal comment is enough.
14:19:00 <chimezie> bijan: There is some fatigue there.  Want to make it light-weight.
14:19:41 <chimezie> ... not sure how D-entailment would help or what the substantive impact is regarding the suggested changes
14:19:50 <chimezie> ... want to understand the changes well enough
14:19:51 <SteveH> q+
14:19:53 <AndyS> q+ to ask about the results format
14:19:56 <LeeF> ack bijan
14:20:00 <LeeF> ack SteveH
14:20:05 <iv_an_ru> yes
14:20:27 <chimezie> SteveH: The changes are substantitive.  It will probably cause another last call
14:20:30 <LeeF> ack AndyS
14:20:30 <Zakim> AndyS, you wanted to ask about the results format
14:20:38 <chimezie> AndyS: Trying to avoid any chance of a last call for rdf:text
14:20:49 <bijan> Hurrah!
14:20:55 <SteveH> that's good
14:21:15 <chimezie> ... there should be a section specifically on SPARQL added
14:21:29 <chimezie> ... wouldn't be unhappy about framing as clarification
14:21:30 <SteveH> if that's true, agreed
14:21:35 <SteveH> but it wasn't my understanding
14:21:47 <AndyS> I don't see that Axel's proposed change addresses the Q's on functions.
14:21:57 <chimezie> AxelPolleres: Change from D-entailment to equivalence.  Don't think it implies a new last call
14:22:48 <AxelPolleres> q+ suggestion
14:22:54 <AndyS> Change is semantic equive to D-entailment (note there is notone singleD-entailment)
14:23:11 <LeeF> ack AxelPolleres
14:23:37 <chimezie> AxelPolleres: Fine with putting suggestion on Wiki page and decide next week ?
14:23:53 <chimezie> AndyS: Would appreciate replies to my email (which included examples)
14:24:22 <chimezie> ... Fine with putting it on Wiki page but it is not subsititute for discussion around the issue(s)
14:25:03 <kjetil> q?
14:25:09 <LeeF> ack suggestion
14:25:13 <chimezie> ... We still aren't discussing result set format
14:25:24 <chimezie> LeeF: AxelP can you take a look at this?
14:25:27 <bijan> q+
14:25:30 <chimezie> AxelPolleres: Sure
14:25:31 <AndyS> ack AndyS
14:25:36 <LeeF> ack bijan
14:26:07 <chimezie> bijan: It might be the case that wee disallow rdf:text in results.  Some entailment regimes might want to do differently
14:26:32 <chimezie> AndyS: It is viable to say it is analagous to RDF graph exchange
14:26:58 <chimezie> bijan: if rdf:text takes off, we will have to revise anyways at some point
14:27:34 <LeeF> topic: face to face
<LeeF> summary: Face-to-face is next Wednesday and Thursday, one hour earlier than planned. 7-3 EDT and 12-8 Bristol UK time
14:27:45 <LeeF> zakim, who's here?
14:27:45 <Zakim> On the phone I see AlexPassant, john-l, AndyS, kasei (muted), SimonS, [Garlik], bijan, AxelPolleres, pgearon, LeeF, Chimezie_Ogbuji (muted), ivanh, kjetil (muted), iv_an_ru, JanneS,
14:27:48 <Zakim> ... PrateekJain-WSU, dnewman2
14:27:48 <Zakim> [Garlik] has SteveH, LukeWM
14:27:49 <Zakim> On IRC I see dnewman2, JanneS, chimezie, LukeWM, SteveH, pgearon, Prateek, Zakim, RRSAgent, kasei, bijan, AndyS, AxelPolleres, iv_an_ru, SimonS, AndyS_, ivanh, kjetil, LeeF,
14:27:52 <Zakim> ... AlexPassant, john-l, KjetilK, ericP, trackbot
14:27:53 <chimezie> LeeF: Confirm attendance (in person versus on phone)
14:27:57 <LeeF> wiki page
14:28:24 <kjetil> Zakim, unmute me
14:28:24 <Zakim> kjetil should no longer be muted
14:28:31 <chimezie> Zakim, unumte me
14:28:31 <Zakim> I don't understand 'unumte me', chimezie
14:28:37 <kasei> I'll be in cambridge next week.
14:28:43 <chimezie> Zakim, unmute me
14:28:43 <Zakim> Chimezie_Ogbuji should no longer be muted
14:28:54 <chimezie> I will not be there in person, *may* be able to participate on phone (not sure)
14:28:56 <Zakim> -kasei
14:29:36 <chimezie> Zakim, mute me
14:29:36 <Zakim> Chimezie_Ogbuji should now be muted
14:30:46 <chimezie> LeeF: Timing of F2F.  
14:30:57 <chimezie> ... an hour earlier?
14:31:47 <chimezie> ... Let us make it an hour earlier
14:32:27 <LeeF> face to face will be 7 - 3 EDT 12 - 8 Bristol time, break each day at 11:30 ET
14:32:53 <chimezie> LeeF: I put on the agenda a rough goal for the F2F
14:33:09 <chimezie> ... don't want to spend the entire time debating features but to begin the deep dive
14:33:30 <kjetil> Zakim, mute me
14:33:30 <Zakim> kjetil should now be muted
14:33:39 <chimezie> ... Perhaps we split the 2 days into 4 half-day blocks.  In one of those we discuss deliverables, etc.
14:33:56 <chimezie> ... in the other 3, start diving into features we have consensus on already (by today possibly)
14:34:49 <chimezie> ... This way we can hit the ground running (WRT to features and our process)
14:35:37 <chimezie> ... Also want to discuss a naming convention
14:35:43 <chimezie> ... input/feedback?
14:35:47 <SteveH> +1, happy
14:35:50 <chimezie> Sounds like a reasonable agenda to me
14:35:54 <AndyS> +1
14:35:54 <LukeWM> sounds ok to me
14:35:56 <pgearon> +1
14:35:57 <kjetil> +1
14:35:57 <Prateek> +1
14:35:58 <AxelPolleres> +1
14:35:59 <AlexPassant> +1
14:36:06 <iv_an_ru> +1
14:36:07 <bijan> +1 to the organizational majesty of lee
14:36:07 <SimonS> +1
14:36:27 <chimezie> LeeF: Questions about logistics?
14:36:32 <kjetil> Zakim, unmute me
14:36:32 <Zakim> kjetil should no longer be muted
14:36:50 <JanneS> q+
14:36:51 <chimezie> kjetil: Car pool from HP labs?
14:37:03 <AndyS> s/from/to
14:37:17 <chimezie> SteveH: I'm going by car, but it is quite small
14:37:28 <iv_an_ru> I'm sorry, I had to escape right now.
14:37:33 <Zakim> -iv_an_ru
14:38:08 <kjetil> Zakim, unmute me
14:38:08 <Zakim> kjetil was not muted, kjetil
14:38:13 <kjetil> Zakim, mute me
14:38:13 <Zakim> kjetil should now be muted
14:38:15 <AndyS> Suggest aim to be at HPL 11:30am for setup.
14:38:16 <LeeF> ack JanneS
14:38:38 <chimezie> JanneS: is HP providing a teleconference #?
14:39:09 <AxelPolleres> should be the same teleconf facility... to be clarified by eric, I guess.
14:39:20 <ivanh> q+
14:39:43 <AxelPolleres> andy: normal phone in the room as "fallback"
14:39:53 <chimezie> ... (conversation continues regarding logistics) ...
14:40:03 <LeeF> ack ivanh
14:40:28 <chimezie> ivanh: We should not have a problem with Zakim
14:40:28 <LeeF> ACTION: LeeF to work with ivanh, ericP to reserve zakim for face to face days
14:40:28 <trackbot> Created ACTION-11 - Work with ivanh, ericP to reserve zakim for face to face days [on Lee Feigenbaum - due 2009-05-05].
14:40:39 <LeeF> ACTION: LeeF to work with EricP to procure a phone for the Cambridge, MA location
14:40:39 <trackbot> Created ACTION-12 - Work with EricP to procure a phone for the Cambridge, MA location [on Lee Feigenbaum - due 2009-05-05].
14:40:41 <AxelPolleres> we shall request zakim for both days, eric should get the physical phone, that should work.
14:40:54 <chimezie> ... prefer to use/have Zakim
14:41:02 <LeeF> topic: feature survey
14:41:16 <LeeF> feature survey -
14:41:31 <LeeF> feature survey results -
14:41:49 <SteveH> q+ to ask about public/private-ness
14:41:53 <SteveH> q-
14:42:03 <chimezie> I wasn't able to decipher priority order from that page
14:42:20 <chimezie> LeeF: Inclincation to have votes on this survey be public, but wanted to run it by group 1st
14:42:24 <chimezie> ... concerns?
14:42:26 <kjetil> +1 on public
14:42:36 <SteveH> +1 from me
14:42:40 <AxelPolleres> PROPOSED: have votes on featue survey be public
14:42:50 <SteveH> raw data:
14:42:50 <AndyS> chimezie, it was opaque with new results it's clear - may need to reload browser (I had caching weirdnesses)
14:42:59 <AxelPolleres> RESOLVED: have votes on feature survey be public
14:43:05 <iv_an_ru> +1 for public votes
14:43:26 <AxelPolleres>
14:43:45 <AndyS> Does the TTL record the "don't wants"?
14:43:59 <chimezie> SteveH: Rendering of another file.  
14:44:05 <SteveH>
14:44:36 <chimezie> ... Process of taking the .ttl page, ran an algorithm to produce a graph
14:44:56 <chimezie> ... people should verify validity.  The javascript is visible as well
14:45:11 <chimezie> ... SVG file captures the Condorcet "beats" relationships
14:45:37 <LeeF> q?
14:45:54 <chimezie> ... Condorcet method looks for pair-wise comparison of every feature voted for and counts how many times feature appear in each vote
14:45:54 <AxelPolleres>
14:46:32 <chimezie> ... can get loops ..
14:47:04 <chimezie> LeeF: one way of interpreting results
14:47:14 <chimezie> ... this was suggested as a good way to look at this
14:47:27 <chimezie> ... it is not quite cut and dry as appears in diagram
14:47:47 <chimezie> ... do we have consensus on the popular features?  For example, Agg functions are at the top
14:48:58 <SteveH> note, my condorcet code doesn't take don't want into account, treats it as won't car
14:49:01 <chimezie> ... AggregateFuncs & Update are the two with consensus importance
14:49:30 <LeeF>
14:49:56 <chimezie> ... Subselects are not top choice, but 11 WG members put it within top 10 ranking.  6 WG members put it in top 3
14:50:15 <chimezie> LeeF: Propose there is consensus on these 3 at least
14:50:46 <chimezie> I agree that these seem to represent consensus
14:50:46 <pgearon> +1
14:50:47 <SteveH> I support that idea
14:50:48 <kjetil> Zakim, unmute me
14:50:48 <Zakim> kjetil should no longer be muted
14:50:48 <AlexPassant> +1
14:50:50 <Prateek> +1
14:50:51 <ivanh> +1
14:50:52 <AndyS> I support these features
14:50:53 <kjetil> +1
14:50:54 <AxelPolleres> +1
14:50:55 <LukeWM> me too
14:51:58 <chimezie> LeeF: Hold off on making them requirements for further conversation
14:52:20 <AndyS> Agree not going to "required" until results are fully in.
14:52:38 <LeeF> PROPOSED: The SPARQL WG adopt AggregateFunctions, Subselects, and Update as three deliverables (either required or time-permitting) for the second phase of the group 
14:52:42 <SteveH> seconded
14:52:50 <kjetil> +1
14:52:55 <AndyS> +1
14:52:58 <ivanh> +1
14:52:58 <AxelPolleres> +1
14:52:59 <Prateek> +1
14:52:59 <JanneS> +1
14:53:01 <pgearon> +1
14:53:03 <SimonS> +1
14:53:14 <LeeF> RESOLVED: The SPARQL WG adopt AggregateFunctions, Subselects, and Update as three deliverables (either required or time-permitting) for the second phase of the group 
14:53:56 <kjetil> Zakim, mute me
14:53:56 <Zakim> kjetil should now be muted
14:54:06 <chimezie> LeeF: would like to (for F2F) solicit volunteers for more design details.  If you can give survey on designs, please let the Chairs know
14:55:29 <chimezie> LeeF: discussion on features with little support but for which we have champions who should discuss why or why not these should be included
14:55:34 <kjetil> Zakim, unmute me
14:55:34 <Zakim> kjetil should no longer be muted
14:56:31 <chimezie> ... there was discussion that limit per resource could be handled by subselects (which is currently a high-priority feature)
14:57:17 <chimezie> kjetil: Now with subselects it is a small thing to implement.  It is extremely important feature because alot of cases we want to list the resources and limit them so we don't get too many solutions
14:57:31 <SteveH> q+ to talk abut FOAF
14:57:33 <chimezie> ... The main selling point is that RDF is suited for heterogenous data
14:58:00 <LeeF> ack SteveH
14:58:01 <chimezie> ... would like to hear from those who don't want it
14:58:02 <Zakim> SteveH, you wanted to talk abut FOAF
14:58:32 <chimezie> SteveH: This is a feature that is needed quite a bit with FOAF in exactly this case: trying to find up to 3 foaf:names and don't want to be overwhelmed with results
14:58:36 <SimonS> q+ to say I am fine with subselects, but do not want explicit syntax
14:58:45 <chimezie> ... prefer subselect aggregate behavior rather than specific syntax in case we get it wrong
14:59:46 <chimezie> Kjetil: Time-permitting feature anyways.  The ORGs that need it, if we do the whole work to specify the syntax, would that be acceptable to WG?
15:00:03 <SteveH> my org needs it for frefernce
15:00:28 <AxelPolleres> q+ to ask about surface syntax
15:00:29 <chimezie> LeeF: Concerned with org-specific spec'ing
15:01:28 <chimezie> ... I feel it would be wise to wait and see if impl. add syntax even with the other features 
15:01:34 <LeeF> ack SimonS
15:01:34 <Zakim> SimonS, you wanted to say I am fine with subselects, but do not want explicit syntax
15:01:35 <chimezie> ... that would make a stronger case for standardization
15:01:54 <chimezie> SimonS: I don't like introducing specific syntax, but happy with subselects addressing this issue
15:01:56 <SteveH> +1 to SimonS 
15:02:06 <chimezie> ... makes sense to build into feature we standardize eventually
15:02:17 <LeeF> ack AxelPolleres
15:02:17 <Zakim> AxelPolleres, you wanted to ask about surface syntax
15:02:32 <chimezie> AxelPolleres: Do we want to subsume such things under surface syntax?
15:02:50 <chimezie> ... avote against surface syntax speaks generally about things like this
15:03:18 <AndyS> SurfaceSyntax is a bit of a catch-all.  I'm wary of putting too much in it.
15:03:28 <chimezie> LeeF: There doesn't seem to be overwhelming support
15:04:02 <chimezie> kjetil: We might want to come back to this after discussion on surface syntax
15:04:38 <chimezie> LeeF: if we accept surface syntax we need a strict definiition 
15:04:59 <chimezie> ... to me it is any feature that can be re-written with identical semantics w/out the new syntax
15:05:52 <Zakim> -kjetil
15:05:58 <chimezie> LeeF: SPARQL/OWL and Parameterized Inference there is confusion on how they are related
15:06:41 <chimezie> bijan: We have extensibility point on semantics of additional answers.  We have a number of implementations
15:07:02 <chimezie> ... it would be nice to converge on SPARQL syntax / semantics.  Have alot of users who want it that move to higher expressivity
15:07:29 <chimezie> ... a separate document and can envision more regimes , so this can be a 'starter'
15:07:32 <AxelPolleres> Bijan, could you paste a link to the BGP extension proposed?
15:07:52 <chimezie> ... it's relation to parameterized inference is that it gives us more regimes to parameterized.  Don't need p-inference to make use of SPARQL/OWL.  
15:08:05 <chimezie> ... independent on how you indicate semantics
15:08:28 <chimezie> LeeF: It would be helpful to address priority
15:08:30 <pgearon> +q
15:08:36 <chimezie> ... there are 2 ORGs that don't want this
15:08:45 <LeeF> ack pgearon
15:08:58 <AxelPolleres> q+ to further explain disambiguation of ParamInference and SPARQL/OWL
15:09:13 <chimezie> pgearon: Don't want to see it, because it could bring the server to its knees
15:09:26 <chimezie> LeeF: This would not be compulsory (and it's own document)
15:09:32 <AndyS> q+ to ask about WG Notes
15:09:39 <ivanh> q+
15:09:40 <chimezie> bijan: compulsory only for systems that *want* to support OWL sensitive query
15:10:02 <LukeWM> q+
15:10:03 <bijan> s/to support OWL sensitive query/
15:10:13 <chimezie> AxelPolleres: I agree with Bijan.  They are complementary.  SPARQL/OWL discusses one entailmemnt regime, p-inference is about requesting (in a query) for a certain entailment regime
15:10:33 <bijan> RDFS should fall out of it, yes
15:10:42 <chimezie> ... do we want to increment up from simple entailment (RDFS, etc..)
15:11:08 <chimezie> ... do we want the WG to work on whether it is requested , do we want the WG to specify advertization of entailment regime (service description)
15:11:13 <LeeF> ack AndyS
15:11:13 <Zakim> AndyS, you wanted to ask about WG Notes
15:11:18 <LeeF> ack AxelPolleres
15:11:18 <Zakim> AxelPolleres, you wanted to further explain disambiguation of ParamInference and SPARQL/OWL
15:11:42 <chimezie> AndyS: WG notes? Not confortable with WG members doing paralle work being injected later into REC track
15:11:56 <pgearon> +1 for the WG to specify advertising entailment regime (and optional features in general)
15:12:06 <chimezie> ... nervous about 'compulsory' implies tests to distinguish if services meet levels of compliance
15:12:19 <ChrisW> ChrisW has joined #sparql
15:13:12 <AndyS> AndyS: Was picking up on the point Lee made about REC track.
15:13:17 <chimezie> LeeF: great point.  Something like SPARQL/OWL (which is orthogonal to main query language) is appropriate for a specification that is reviewed as a REC track or Note
15:13:30 <chimezie> ... SPARQL/OWL is appropriate for this 
15:13:36 <chimezie> bijan: Was thinking about REC track
15:14:11 <chimezie> ... I have been tasked to write this up anyways.  Want to force CR on implementations to force convergence 
15:14:16 <JanneS> (sorry gotta run home) - hear you next week
15:14:26 <Zakim> -JanneS
15:14:37 <chimezie> AndyS: If we put  it on REC track we have a finite amount of time to discuss issues
15:14:47 <chimezie> bijan: Agreed.  We don't have to decide until late in the game
15:15:11 <chimezie> ... willing to do work with possible outcome of a Note.  If we can make it with reasonable amount of effort, it wouldn't preclude a REC doc
15:15:21 <chimezie> ... a Note is better than nothing
15:15:28 <LeeF> q?
15:16:19 <AxelPolleres> q+
15:16:20 <chimezie> ivanh: do we plan to do anythign at all with rules.  If everything goes as planned, by the time this SPARQL is a REC we ill have RIF as a REC
15:16:30 <AndyS> If SPARQL/OWL, be great for a RDFS (RDFS++) Note as well.
15:16:33 <chimezie> ... something should say how SPARQL relates to RIF.. we should be careful 
15:16:45 <chimezie> +1 with ivanh about isolating SPARQL from other standards
15:17:00 <pgearon> q+
15:17:09 <chimezie> Zakum unmute me
15:17:14 <LeeF> ack ivanh
15:17:16 <chimezie> Zakum, unmute me
15:17:31 <chimezie> Parematerized inference does give an extension point to RIF as described in the Wiki
15:17:40 <LeeF> chimezie, thanks, you are right
15:17:48 <AndyS> zakim, unmute chimezie
15:17:49 <Zakim> Chimezie_Ogbuji should no longer be muted
15:17:55 <SteveH> but arguable not an appropriate one...
15:18:11 <AxelPolleres> +1 to ivanh
15:18:16 <LeeF> q?
15:18:20 <LeeF> ack LukeWM
15:18:31 <chimezie> zakim, mute me
15:18:31 <Zakim> Chimezie_Ogbuji should now be muted
15:18:47 <chimezie> LukeWM: don't know that much about OWL, haven't had much experience with it, mostly a matter of priority
15:18:59 <LeeF> ack AxelPolleres
15:19:23 <chimezie> AxelPolleres: RIF+RDF graphs complies well with entailment regimes for SPARQL
15:20:19 <chimezie> ... they should go together
15:20:24 <bijan> q+ to ask if this is how we mean to go on as opposed to the language groups doing them themselves
15:20:27 <chimezie> LeeF: Time is the primary caveat
15:20:32 <SteveH> I suspect that RIF and OWL raise slightly different issues
15:20:44 <bijan> SteveH: for sure
15:21:03 <chimezie> ... doing one of these will be a good way to test the current extension point
15:21:15 <LeeF> ack pgearon
15:21:54 <chimezie> pgearon: Rules are: rule-based query and rulesets that generate statements (falls into area of update)
15:22:25 <chimezie> zakim, unmute me
15:22:25 <Zakim> Chimezie_Ogbuji should no longer be muted
15:22:48 <LeeF> chimezie: I think there is an overlap between what SPARQL/OWL and parameterizedinference are tryingto achieve
15:22:54 <LeeF> ... important to spell out more clearly than we currently do
15:23:16 <LeeF> ... better to do one or the other than nother
15:23:29 <LeeF> ... important as we do more expressive querying of the semantic web
15:23:46 <SteveH> I don't like conflating ParameterisedInference and SPARQL/OWL
15:23:50 <LeeF> ack bijan
15:23:50 <Zakim> bijan, you wanted to ask if this is how we mean to go on as opposed to the language groups doing them themselves
15:23:58 <SteveH> they are really quite different
15:24:18 <chimezie> bijan: perhaps WGs and community can do this
15:24:24 <ivanh> q+
15:24:32 <SimonS> SteveH +1
15:24:36 <chimezie> This argument applies to OWL as well as RIF , BTW (having this done in separate communities)
15:25:00 <ivanh> q-
15:25:05 <AxelPolleres> +1 to that OWL and RIF did already a lot of pre-work in these regards and it shouldn't be so difficult as some expect.
15:25:30 <ivanh> q+
15:25:34 <chimezie> SteveH: one is about querying store with inference capability and the other is about for this query , use this set of features
15:25:52 <chimezie> The difference is that in once case you are being specific about the 'feature' in the other, the feature is open ended
15:25:56 <chimezie> zakim, unmute me
15:25:56 <Zakim> Chimezie_Ogbuji was not muted, chimezie
15:26:43 <SteveH> q+ (again sorry)
15:26:48 <LeeF> chimezie: RDF clinical data - want to derive variables for reporting to external agency - reporting requirements come after the fact - makes sense to write constraints for how to derive variables and then include that as parameter to the query
15:28:04 <chimezie> ivanh: if we forget RIF for a moment, we have various ways to add inference (RDF, OWL2, RDFS, etc..)
15:28:27 <chimezie> ... a finite list. Not really parameterized.  With RIF we have something else.  A well0defined way to define rules 
15:29:09 <SimonS> q+ the (RDF, OWL2, ...) case is not that different if you consider ontologies as parameters analogous to rulesets
15:29:28 <LeeF> ack SteveH
15:29:31 <LeeF> ack ivanh
15:29:34 <SimonS> q+ 
15:29:37 <chimezie> SteveH: different point.  Prior was working on rule-based query engines.
15:29:45 <chimezie> ... didn't require syntax extension
15:30:09 <Zakim> -john-l
15:30:13 <chimezie> ... the concern is regarding the proposed syntax.  Doesn't cover this usecase  .  Use services instead of rules
15:30:27 <chimezie> ... the community therefor doesn't have enough consensus
15:30:40 <bijan> It seems that people are confusing defining an entailment/generation regime with the task of assigning such a regime to a particular BGP
15:31:19 <AxelPolleres> q+ why just FROM is not so easy.
15:31:29 <chimezie> the latter is needed in either case
15:31:32 <AxelPolleres> q?
15:31:35 <AxelPolleres> q+
15:31:46 <AndyS> +1 to SteveH - need parts of query, not just overall
15:31:52 <bijan> Why a rule set instead of an arbitrary extra graph?
15:32:07 <chimezie> i.e., specifying a regime is only useful if you can 'use' it explicitely.  I think of this latter part as the common ground between both
15:32:43 <bijan> chimezie, sure (though you can do it with endpoints), but they are distinct.
15:33:55 <chimezie> LeeF: not sure where are .  Perhaps continue on mailing list?
15:34:00 <chimezie> ... pick it up during F2F
15:34:04 <chimezie> ... adjorn for today