16:52:21 RRSAgent has joined #owl 16:52:21 logging to http://www.w3.org/2008/09/10-owl-irc 16:52:34 RRSAgent, make records public 16:56:42 SW_OWL()1:00PM has now started 16:56:49 + +0190827aaaa 16:57:01 zakim, aaaa is me 16:57:01 +MartinD; got it 16:57:21 IanH has joined #owl 16:58:00 +Sandro 16:58:02 MartinD has changed the topic to: http://www.w3.org/2007/OWL/wiki/Teleconference.2008.09.10/Agenda 16:58:36 +Ian_Horrocks 16:58:43 bmotik has joined #owl 16:58:51 zakim, Ian_Horrocks is IanH 16:58:51 +IanH; got it 16:58:52 Zakim, this will be OWL 16:58:53 ok, bmotik, I see SW_OWL()1:00PM already started 16:59:24 RRSAgent, make records public 16:59:28 +??P6 16:59:31 Zakim, ??P6 is me 16:59:31 +bmotik; got it 16:59:34 Zakim, mute me 16:59:34 bmotik should now be muted 16:59:34 zakim, who is here? 16:59:35 On the phone I see MartinD, Sandro, IanH, bmotik (muted) 16:59:36 On IRC I see bmotik, IanH, RRSAgent, Zakim, MartinD, baojie, sandro, alanr, trackbot 16:59:59 Martin, are you all set for scribing? 17:00:06 hope so... :-) 17:00:14 m_schnei has joined #owl 17:00:16 zakim, mute me 17:00:16 MartinD should now be muted 17:00:20 bcuencagrau has joined #owl 17:00:29 Zhe has joined #owl 17:00:40 zakim, who is here? 17:00:40 On the phone I see MartinD (muted), Sandro, IanH, bmotik (muted) 17:00:41 On IRC I see Zhe, bcuencagrau, m_schnei, bmotik, IanH, RRSAgent, Zakim, MartinD, baojie, sandro, alanr, trackbot 17:00:42 uli has joined #owl 17:01:00 + +1.603.897.aabb 17:01:12 zakim, +1.603.897.aabb is me 17:01:15 +??P13 17:01:19 +Zhe; got it 17:01:20 Achille has joined #owl 17:01:22 zakim, ??P13 is me 17:01:23 zakim, mute me 17:01:27 +m_schnei; got it 17:01:29 Zhe should now be muted 17:01:32 +[IBM] 17:01:37 Zakim, IBM is me 17:01:37 +Achille; got it 17:01:41 +??P14 17:01:48 zakim, ??P14 is me 17:01:48 +uli; got it 17:01:52 zakim, mute me 17:01:52 uli should now be muted 17:01:58 +??P16 17:02:00 zakim, mute me 17:02:00 m_schnei should now be muted 17:02:05 Zakim, ??P16 is me 17:02:05 +bcuencagrau; got it 17:02:11 Zakim, mute me 17:02:11 bcuencagrau should now be muted 17:02:14 zakim, who is here? 17:02:14 On the phone I see MartinD (muted), Sandro, IanH, bmotik (muted), Zhe (muted), m_schnei (muted), Achille, uli (muted), bcuencagrau (muted) 17:02:16 On IRC I see Achille, uli, Zhe, bcuencagrau, m_schnei, bmotik, IanH, RRSAgent, Zakim, MartinD, baojie, sandro, alanr, trackbot 17:02:35 IanH: let us start with today's agenda 17:02:45 Topic: Admin 17:02:58 + +1.518.276.aacc 17:03:00 IanH: any agenda amendments 17:03:15 Zakim, aacc is baojie 17:03:15 +baojie; got it 17:03:29 IanH: Previous minutes (http://www.w3.org/2007/OWL/meeting/2008-09-03) 17:04:04 PROPOSED: Accept Previous Minutes (3 September) 17:04:07 +1 17:04:12 +1 17:04:15 +1 17:04:22 +1 ;) 17:04:34 RESOLVED: Accepted Previous Minutes (3 September) 17:04:47 Subtopic: Pending actions 17:05:01 +Alan 17:05:09 IanH: usual procedure, let's see how actions were completed, people may say why not completed 17:05:21 IanH: if no objections, we assume actions are done... 17:05:26 189 not done 17:05:31 he did 17:05:42 IanH: Action 179 seems to be complete 17:05:55 q? 17:06:15 Ianh: Action 172 - achille suggests next Tuesday as a day to complete the action 17:06:24 q? 17:06:33 IanH: Action 189 - Alan says this is not done 17:06:46 Alanr: action 189 should be next week 17:07:01 InaH: Action 185 - should be done, if I remember correctly 17:07:17 ...yes, it is done 17:07:28 msmith has joined #owl 17:07:28 q? 17:07:29 ... Action 202 - was on Alan 17:07:53 AlanR: still pending, will provide update in the near future 17:07:53 zhe also finished 17:07:58 yes 17:08:07 ianH: Action 181 done by Zhe 17:08:09 ... 17:08:16 q? 17:08:51 + +1.202.408.aadd 17:08:56 Sandro: Action 207, publication plan (as created last week) - join pub by RIF and OWL groups? 17:08:56 q? 17:09:15 ... this action should be made a bit clearer 17:09:57 IanH: last week we agreed rough plan how this publication can happen and there is an action on how this should be implemented 17:10:04 q? 17:10:07 sandro, the context is at http://www.w3.org/2007/OWL/meeting/2008-09-03#Pending_actions 17:10:11 ... probably this week's deadline was a bit optimistic 17:10:46 Sandro: apparently, a joint recommendation is a good thing, if it can be achieved 17:10:49 q? 17:11:03 ... there need to be two resolutions to publish (from two groups) 17:11:36 IanH: if Sandro is the contact on both, it might be good to watch that the process is moving ahead, a kind of monitoring 17:11:51 ... we will fix the action text later 17:12:01 q? 17:12:04 ... Action 174 is on Bijan 17:12:06 zakim, who is here? 17:12:06 On the phone I see MartinD (muted), Sandro, IanH, bmotik (muted), Zhe (muted), m_schnei (muted), Achille, uli (muted), bcuencagrau (muted), baojie, Alan, msmith 17:12:09 On IRC I see msmith, Achille, uli, Zhe, bcuencagrau, m_schnei, bmotik, IanH, RRSAgent, Zakim, MartinD, baojie, sandro, alanr, trackbot 17:12:21 ...i will go down the corridor and knock... 17:12:23 ewallace has joined #owl 17:12:34 ... no Bijan yet, so we need to check later what is the status of this action 17:12:51 Subtopic: Reviewing 17:13:08 IanH: thank you to all who reviewed documents and gave feedback, good job! 17:13:24 q? 17:13:32 ... one exception is the Profile - not a fault of reviewers, but there is still some discussion ongoing 17:13:40 ... hope to conclude this within few days 17:13:41 bijan has joined #owl 17:14:03 ... according to the schedule from F2F meeting, we should publish the drafts by September 15... 17:14:04 q+ 17:14:06 +Peter_Patel-Schneider 17:14:09 q? 17:14:13 zakim, unmute me 17:14:13 m_schnei should no longer be muted 17:14:18 pfps has joined #owl 17:14:20 ... perhaps people working on the docs may say if this is still realistic 17:14:21 q? 17:14:55 I'm nowhere near done my review, but I'm comfortable publishing without it (Syntax is a big document!) 17:14:55 q? 17:15:04 zakim, mute me 17:15:04 m_schnei should now be muted 17:15:05 m_schnei: let's wait for the next stage, in my case we will finish the review by Friday... but there will be some potential points that may need further discussion 17:15:08 q? 17:15:17 q- 17:15:21 IanH: we can wait a few days to give people time to review things properly 17:15:36 ... any objections to delaying the publication by a few days? 17:15:44 q? 17:15:46 I'll try to handle the reviews of Syntax this weekend 17:15:58 ... what about syntax? do we have a doc that reflects reviews by next week 17:16:00 it's done. 17:16:03 Dnoe 17:16:10 ... model theoretic semantics is done 17:16:13 q? 17:16:17 ... what about RDF? 17:16:24 essentially done, needs a little bit more work 17:16:38 yes, I expect to be done later today 17:16:42 ... is it realistic to publish it next week? 17:16:42 q? 17:17:00 q+ 17:17:05 q? 17:17:09 ack pfps 17:17:09 Sandro: is there some proposal in there on importing? 17:17:34 that wasn't me, MartinD 17:17:42 -Alan 17:17:47 q? 17:18:02 IanH: we still have some open issues, there will be editorial comments that would clarify parts that can change 17:18:06 I think it's done 17:18:09 ... xml serialization? 17:18:09 done 17:18:18 ... review of this doc is done 17:18:32 q+ 17:18:36 q? 17:18:38 ... we're in good shape, so we should be in position to vote on publication of these docs next week 17:18:41 ack pfps 17:19:02 pfps: those people who did reviews should perhaps check that their comments are adequately reslved/addressed 17:19:18 q? 17:19:35 +Alan 17:19:39 IanH: typically, these reaction are happening, but reviewers should perhaps check that this is really happening 17:19:40 q? 17:19:51 q? 17:20:37 q? 17:20:49 ... when editors finish updates according to the reviews, we should send a msg to WG mailing list to alert people who want to re-check... 17:20:59 ... so that we can hold the vote next week 17:21:07 q? 17:21:19 ... editors should let Ian know about the status 17:21:22 yes 17:21:29 ... all happy with doc publication 17:21:41 Subtopic: SKOS last call draft 17:21:43 q+ 17:21:48 q? 17:21:53 ack pfps 17:21:55 IanH: no volunteers last week to review it, so still on agenda 17:22:02 pfps: there is a review by me... 17:22:22 ... not quite sure what to do with my review, but it might act as a basis for WG review 17:22:25 q+ RIF Review for OWL 2 17:22:25 goal would be to see what can/can't be represented in owl2 17:22:28 q? 17:22:31 q+ to ask about RIF Review for OWL 2 17:22:32 ... there are more than one document 17:22:38 q? 17:22:38 AFAIK, only the SKOS reference is in Last Call 17:22:38 IanH: any volunteers? 17:22:53 q? 17:23:01 ack sandro 17:23:01 sandro, you wanted to ask about RIF Review for OWL 2 17:23:05 q? 17:23:11 I'm working on my own review (work in progress) 17:23:30 zakim, unmute me 17:23:30 m_schnei should no longer be muted 17:23:30 q? 17:23:37 ... can Jie perhaps check if someone from there wouldn't do it 17:23:38 q? 17:23:58 m_schnei: working on a review, but not sure if there should be an "OWL WG" official 17:24:14 q+ 17:24:17 q? 17:24:21 ack alanr 17:24:23 IanH: if Peter and Michael finish their reviews, we may consider them both and discuss (if needed) 17:24:48 Alan: what aspects are you focusing on? e.g. to what extent SKOS relates to OWL profile(s) 17:25:00 pfps: this has been partly done, details to follow later 17:25:18 m_schnei: i'm interrested in RDF semantics and those factors 17:25:30 q? 17:25:47 zakim, mute me 17:25:47 m_schnei should now be muted 17:25:49 Alan: if you are willing to contribute your reviews, we can see if we agree on a common statement/review 17:26:03 IanH: let's see what comes from Peter and Michael 17:26:10 Subtopic: F2F meeting 17:26:23 IanH: indicate your status on the page on the wiki 17:26:37 q? 17:26:40 ... http://www.w3.org/2007/OWL/wiki/F2F4 17:26:54 Sandro: suggestion for agenda amendment 17:27:16 actually, I helped write it, so I"m not sure that I *reviewed* it 17:27:38 ... RIF doc review was done mostly with OWL 1 focus, maybe there can be a check on whether OWL WG is still happy with it 17:27:41 q? 17:27:49 at first blush, I can't think of any changes required (but don't let me bias the review) :-) 17:27:50 ... someone other than Peter who helped writing it 17:28:09 http://www.w3.org/TR/rif-rdf-owl/ 17:28:14 q? 17:28:29 are there timelines? 17:28:46 q? 17:29:02 Sandro: it's about next few days, so it may be a bit tough to do it within deadlines 17:29:24 IanH: not many people volunteering, perhaps we need an email to reach to other people 17:29:45 ... admin concluded 17:29:49 Topic: Issues 17:29:58 IanH: two resolution proposals 17:30:01 q? 17:30:09 q+ 17:30:12 ... issue 133 on DL-Lite profile 17:30:25 zakim, who is on the call? 17:30:25 On the phone I see MartinD (muted), Sandro, IanH, bmotik (muted), Zhe (muted), m_schnei (muted), Achille, uli (muted), bcuencagrau (muted), baojie, msmith, Peter_Patel-Schneider, 17:30:28 ... Alan 17:30:39 q? 17:30:43 ack msmith 17:30:46 msmith: the proposal is to move functional property and ... axioms from the profile... 17:31:09 ... there should be a core DL-Lite that does not have all those extensions 17:31:15 +q 17:31:31 IanH: there might be some text in the profile doc mentioning about these exceptions? 17:31:34 q? 17:31:44 Zakim, unmute me 17:31:44 bcuencagrau should no longer be muted 17:31:47 msmith: yes, this should happen and diego was also happy 17:31:51 ack bcuencagrau 17:32:00 bcuencagrau: unclear what was proposed... 17:32:23 q? 17:32:27 ... do we have DL-Lite and then concerning assertions will we still have sameAs and differentFrom? 17:32:51 msmith: differentFrom is acceptable, sameAs probably not 17:33:02 q? 17:33:21 bcuencagrau: we have basic features in the profile 17:33:29 "the intersection" of the choices is how i see it 17:33:34 q? 17:33:52 ... there are only axioms, no unique axioms? 17:34:14 Zakim, mute me 17:34:14 bcuencagrau should now be muted 17:34:17 q? 17:34:18 ... what we have in the doc has been proposed a fe months ago 17:34:28 looks good to me 17:34:42 I am fine with it too 17:34:48 IanH: given there were no objetions in email, we propose to resolve it 17:35:01 PROPOSED: Resolve Issue 133 (# Issue 133 (DL-Lite Profile modified to include UNA) per Mike's email ) 17:35:04 +1 17:35:07 +1 17:35:07 +1 17:35:08 +1 17:35:10 +1 17:35:13 +1 17:35:14 +1 17:35:17 +1 17:35:22 Mike's email = http://lists.w3.org/Archives/Public/public-owl-wg/2008Sep/0017.html 17:35:35 +1 17:35:46 RESOLVED: Issue 133 (DL-Lite Profile modified to include UNA) per Mike's email (http://lists.w3.org/Archives/Public/public-owl-wg/2008Sep/0017.html) 17:36:03 Subtopic: Issue 119 17:36:04 Zakim, mute me 17:36:04 bcuencagrau was already muted, bcuencagrau 17:36:11 q? 17:36:17 IanH: this seems to be resolved by RDF semantics 17:36:34 ... due to self-restriction this could have been a problem, but it was resolved by Mike 17:36:39 ... not really controversial 17:36:42 q? 17:36:58 PROPOSED: Resolve Issue 119 (OWL 2 Full may become inconsistent due to self restrictions) per Ian's email 17:37:03 +1 17:37:06 +1 17:37:09 +1 17:37:09 +1 17:37:10 +1 17:37:12 +1 17:37:14 Ian's email = http://lists.w3.org/Archives/Public/public-owl-wg/2008Sep/0033.html 17:37:17 +1 17:37:18 +1 17:37:25 +1 17:37:29 +1 17:37:36 +1 17:37:45 RESOLVED: Issue 119 (OWL 2 Full may become inconsistent due to self restrictions) per Ian's email (http://lists.w3.org/Archives/Public/public-owl-wg/2008Sep/0033.html) 17:38:02 Subtopic: Other issues requiring discussion 17:38:18 Subsubtopic: Issue 130 - conformance 17:38:31 IanH: this has been discussed last week, a few emails... 17:38:35 q+ 17:38:36 q? 17:38:43 ack sandro 17:38:43 ... spend a few minutes to get a resolution? 17:38:54 Sandro: we exchanged some emails and mostly we're happy 17:39:09 q? 17:39:10 ... there was a proposal to amend some text, I liked it 17:39:29 IanH: shall be make a change agreed in the email? 17:39:32 q+ 17:39:33 make change and produce a proposal 17:39:39 q? 17:39:41 zakim, unmute me 17:39:41 Zhe should no longer be muted 17:39:43 ... ok, let's assume we go for the change 17:39:52 ack Zhe 17:39:54 pointer 17:40:10 q+ 17:40:19 http://lists.w3.org/Archives/Public/public-owl-wg/2008Sep/0029.html 17:40:22 q? 17:40:31 the text starting "An OWL 2 RL...." 17:40:33 IanH: I will update the conformance document with the modified text and I will send an email how was this implemented, so that people can comment 17:40:56 q? 17:40:59 ack alanr 17:41:02 q? 17:41:05 ... proposals from the author re words like could, should,... will be made into the text too 17:41:19 Unknown, Reason= 17:41:19 - Resource Limits Reached 17:41:19 - Finished Incomplete Algorithm 17:41:19 - Unexpected Error 17:41:26 Alan: yesterday we discussed with Sandro - there are two meanings of "unknown" 17:41:37 ... unable to complete, e.g. due to resource limitations 17:41:50 q? 17:41:58 ... another is due to finished but not guaranteed entailment alg. 17:42:22 q? 17:42:27 ... then, if the answer doesn't make sense... we may not have a terminating message 17:42:36 q+ is this a test case question or an API question? 17:42:40 q? 17:42:45 q+ to ask is this a test case question or an API question? 17:42:56 ... a proposal for something that would make clear that alg. ran out of resources vs. not knowing the answer 17:43:09 "Out of Resource" sounds pretty technical for a formal spec 17:43:16 q? 17:43:21 ack sandro 17:43:21 sandro, you wanted to ask is this a test case question or an API question? 17:43:23 ... even if these messages are present in OWL 1, there is no reason why to keep previous language 17:43:37 Sandro: I pasted the three meanings of "unknown" above 17:44:05 ... not sure how useful this is; it can help in test cases, but not sure how valuable this would be in API 17:44:05 {True, False, Unknown} is better than {True,False} in Prolog 17:44:27 I DONT think it helps in the test cases. 17:44:29 q? 17:44:38 q+ 17:44:39 IanH: one can perhaps distinguish even more cases to complement values of true and false 17:44:41 q? 17:44:46 ack alanr 17:44:50 ... any opinions from implementers 17:45:30 q? 17:45:35 ... one case where it makes sense is when the check hasn't been done, so it maybe undesirable to return unknow 17:45:49 "Completed-Unknown" 17:45:54 q? 17:46:21 q? 17:46:22 q+ 17:46:25 Ian: {True, False, UnexpectedError, CompletedComputationButNoAnswer } 17:46:26 zakim, unmute me 17:46:26 m_schnei should no longer be muted 17:46:27 q? 17:46:28 =0 17:46:32 +1 four cases for OWL RL 17:46:36 +0 17:47:03 m_schnei: one can put comments re conformance, e.g. for OWL Full it cannot be avoided that "unknown" will come out 17:47:11 q? 17:47:21 zakim, mute me 17:47:21 m_schnei should now be muted 17:47:33 perhaps we can see the different alternatives in writing? 17:47:39 +1 to Ian 17:47:44 ack m_schnei 17:47:47 IanH: I will have another pass on the doc and see if people like it 17:47:49 q? 17:48:22 Sandro: we should say that in general, one should not be returning "unknown", which may pose a conflict with an OWL test case? 17:48:35 ... what about query answering issues? 17:49:10 q? 17:49:25 IanH: we can mention something like XML query answering and show how these entailment checks would impact on QA... rather than having a complete new sections on QA 17:49:40 Subtopic: Issue 144 17:49:41 q? 17:49:48 q+ 17:49:53 SCRIBE-CORRECTION: No, what I said was that there is nothing wrong with returning "unknown" in OWL RL. 17:49:55 q? 17:49:58 ack Zhe 17:50:05 also note: http://lists.w3.org/Archives/Public/public-owl-comments/2008Jul/0002.html 17:50:12 Zhe: we discussed this in the WG before... 17:50:32 ... if we don't include the base triple we may put unnecessary burden on implementations 17:50:33 q? 17:50:33 q+ 17:50:35 q+ 17:50:39 q+ 17:50:42 Zakim, unmute me 17:50:42 bmotik should no longer be muted 17:50:48 ... we are suggesting to simply including it, which makes life easier 17:50:50 q? 17:50:54 ack bmotik 17:50:55 q+ 17:51:20 Boris: seems like reasonable thing to do but the problem is that an axiom is not represented as one thing vs. two things 17:51:39 ... what if you find both - base axiom and the reified one... then what? 17:52:00 ... we may decide e.g. on forgeting the base one if reified axiom is found 17:52:06 q? 17:52:08 ... this may cause some mapping issues 17:52:35 ... then there is another issue = including the triple does not tell you what to do with it or if it is not find 17:53:00 ... something along lines "from reified triple define the original" 17:53:05 q? 17:53:11 zakim, unmute me 17:53:11 m_schnei was not muted, m_schnei 17:53:23 ... should we start adding original triples if we find a reified one 17:53:39 Zakim, mute me 17:53:39 bcuencagrau was already muted, bcuencagrau 17:53:48 ... finally, I don't think this will occur often enough, so that it can cause problems with efficiency 17:53:54 q? 17:54:09 ack m_schnei 17:54:25 m_schnei: without the triples it seems more stable? 17:54:39 Boris has made my points 17:54:41 q- 17:54:52 ... would current RDF serializations... 17:55:31 ... if it is not always avoidable to have triple in (if you want to annotate the triple without having access to the orig. ontology), would you define new ontology? 17:55:33 q? 17:55:47 ... there might arise problems with axiom closure 17:55:58 ... I would not be in favour, not necessary IMHO 17:56:07 q? 17:56:17 zakim, unmute me 17:56:17 m_schnei was not muted, m_schnei 17:56:22 q+ 17:56:39 Alan: what about missing base triple -- there is a syntax for it, so no major issue 17:57:08 q? 17:57:12 q+ 17:57:13 ... regarding michael's comment, not sure this would be a really problem, perhaps only in some profiles? 17:57:18 ack alanr 17:57:18 of course, you can have two ontology files, one having the spo, the other having the reification, and then having the second import the first 17:57:20 q+ 17:57:23 q+ to ask why Alan's example is monotonic 17:57:31 s/monotonic/nonmonotonic/ 17:57:33 ... issues are not really with performance, more about monotonicity... 17:57:41 ack bmotik 17:58:00 q+ 17:58:27 last statement re OWL RL seems wrong. OWL RL has specific syntax. 17:58:31 Boris: if triple is not there, one can reverse-parse it... but what would OWL-RL parser do with this... if you have RDF graph without this triple, you are missing on some inferences 17:58:43 conformance allows OWL RL entailment checked to take and RDF 17:58:49 s/checked/checker/ 17:58:49 ... there is no guarantee the triple will be included (as it should... 17:58:53 q? 17:59:18 yes, OWL Full infers the spo 17:59:28 q? 17:59:33 ... then about monotonicity, we already have in OWL Full semantics, there is a possibility to get to non-reified version by means of reasoning... 17:59:34 where is there that reification implies base triple? 17:59:36 wasn't in RDF 17:59:57 pfps: don't think Alan's example is non-mononotonic 18:00:00 q? 18:00:03 ack pfps 18:00:03 pfps, you wanted to ask why Alan's example is monotonic 18:00:06 q+ 18:00:08 ack Zhe 18:00:09 Zhe: still want to stress the performance issue 18:00:11 q? 18:00:23 ... if an application wants to use this type of annotation 18:00:49 ... you can imagine this is an additional burden to keep checking on info on every single triple 18:00:51 q+ to do a back-of-the-envelope calculation of the relative costs 18:01:19 ... if base triple is out, it's possible, it's not efficient... if there is a mix of annotated and non-annotated axioms, what should we do? 18:01:20 Zhe, perhaps this can be overcome by some clever data structures? 18:01:26 q? 18:01:37 ... should we accept axiom with annotation and forget the ones without annotation? 18:01:50 q? 18:01:57 ack msmith 18:02:05 msmith: axiom with and without annotation are structurally different 18:02:12 +1 to msmith 18:02:18 q? 18:02:22 ... this is already in the spect 18:02:26 ack bmotik 18:02:51 Boris: we can address the concerns with performance without altering the core 18:03:27 ... people can produce RDF graphs... it is safer to assume that one gets RDF graph that needs checking if things are in it 18:03:40 ... we can think about ways to handle certain common cases 18:03:42 q? 18:04:05 Q: How does RDF semantics 4.18 avoid asserting positive triple for negative property assertion? 18:04:07 ... the biggest problem with reification is their occurrence in different part of file = problem for parsers that need to trace this 18:04:10 q? 18:04:31 ... suggestion: implementation should put reified triples together, one after another... 18:04:40 we don't have control of this in the rdf world 18:04:45 q? 18:04:46 ... this would allow more efficient handling 18:05:27 rdf pipes, etc 18:05:31 ... of course, we don't have any control over this... but OWL things are written in files, so we may recommend it? 18:05:41 q? 18:05:46 ack pfps 18:05:46 pfps, you wanted to do a back-of-the-envelope calculation of the relative costs 18:05:50 q? 18:05:57 Peter: there was a point about performance issue, 18:06:09 ... reading a triple is expensive, even compared to running rules 18:06:19 q? 18:06:21 a whole lot? 1/3 of # axioms that are annotated 18:06:23 http://pipes.deri.org/ 18:06:29 q+ 18:06:40 ... if we had more triples, we are likely to increase the amount of I/O required 18:06:41 q? 18:06:49 Zhe: maybe by 20-30% 18:07:15 q? 18:07:21 Peter: yes, but that's quite substantial... unless we do an actual analysis, I am not prepared to support that we would save actual resources 18:08:03 joins more expensive than io 18:08:05 Zhe: if annotation axioms does not include the base triple, we need to do additional joins in the tables... 18:08:09 q+ 18:08:38 I'm not prepared to admit that in a decent implementation that rule processing is more exensive than adding triples 18:08:43 q? 18:08:47 IanH: hard to establish what takes more time - loading triples into table or doing joins.... 18:09:18 Boris: briefly about RDF pipes... unlikely that you cannot ship related triples 18:09:24 q? 18:09:29 ack bmotik 18:09:37 pipes: not if they go through some hash table as part of their processing 18:09:46 which is likely 18:10:17 anyways, implementation has to handle worse case 18:10:22 ... if we are processing arbitrary RDF graph, if we have guarantees that in reasonable cases the triples would be close, one can implement a thing that would basically read X triples and replace them with the base triple (if that's needed) 18:10:38 q? 18:10:39 -Alan 18:10:44 q? 18:10:56 .... if we make sure the triples are close to each, we can leave the spec as it is, and you have control over your implementation 18:11:22 q? 18:11:29 IanH: what about doing the thing in tables, in a similar way as you said, filling table once? 18:11:35 +Alan 18:11:55 Boris: true but one may actually save on filling and re-filling the table because the axiom comes later? 18:11:59 q? 18:12:02 zakim, unmute me 18:12:02 m_schnei was not muted, m_schnei 18:12:26 IanH: sounds interesting... appropriate to take discussion offline for the interested parties, so that they come up with a proposal to resolve this... 18:12:35 q? 18:12:37 ... ideally by not having to have base triples? 18:12:42 ack m_schnei 18:12:50 m_schnei: I/O is perhaps not interesting 18:13:11 q? 18:13:20 ... if we find the version of the triple but not the original triple... what is *wrong* with this (disregarding I/O performance) 18:13:37 IanH: there is no reverse mapping for OWL Full though 18:13:44 q? 18:13:48 m_schnei: I mean OWL DL 18:13:53 q? 18:14:00 q- 18:14:04 zakim, mute me 18:14:04 m_schnei should now be muted 18:14:14 IanH: but the discussion is now about OWL RL, so ... let's take this offline and see if things are resolved this way 18:14:21 Subtopic: Issue 109 18:14:34 q? 18:14:37 IanH: last time we were close to resolving namespaces in this issue? 18:14:49 q? 18:14:49 ... no conclusions yet 18:15:24 :) 18:15:31 Sandro: we are waiting for getting some objective... we need to find technical differences to rule one way or another 18:16:11 IanH: so at the end of discussion we will somehow need to flip the coin, unless there is an agreement between protagonists 18:16:19 Sandro: do we have pros and cons? 18:16:28 +??P5 18:16:31 q? 18:16:38 IanH: we looked at it from different angles and the point is in different opinions 18:17:01 bijan has joined #owl 18:17:04 Alan: is this an architectural issue? 18:17:16 I won't accept TAG arbitration 18:17:22 ... if this is on stake, why not bringing some else in? 18:17:25 zakim, who is here 18:17:25 bijan, you need to end that query with '?' 18:17:35 zakim, who is here? 18:17:35 On the phone I see MartinD (muted), Sandro, IanH, bmotik, Zhe, m_schnei (muted), Achille, uli (muted), bcuencagrau (muted), baojie, msmith, Peter_Patel-Schneider, Alan, ??P5 18:17:38 On IRC I see bijan, pfps, ewallace, msmith, Achille, uli, Zhe, bcuencagrau, m_schnei, bmotik, IanH, RRSAgent, Zakim, MartinD, baojie, sandro, alanr, trackbot 18:17:44 zakim, ??p5 is me 18:17:44 +bijan; got it 18:17:46 q+ 18:17:51 q? 18:18:12 ... is there a suggestion where we can ask for ideas? e.g. XML WG 18:18:35 I would listen to TAG opinion 18:18:38 Alan: do we need more time to this? next week? 18:18:54 yes 18:18:58 q? 18:19:02 ack bijan 18:19:37 Bijan: curious about these situations, there should be some evidence which we don't have at the moment... mere judgments are not really making much difference here 18:19:53 IanH: in the end there will have to be a vote on this in WG 18:20:52 ... so it's really about other members of WG to make up their minds and in voting go one way or another... so far it's mainly W3C and Manchester objecting (with most being indifferent) 18:20:54 I object to that 18:21:02 q? 18:21:07 ... so what about that coin idea = if no decision reached 18:21:51 ... when do we expect to make this decision 18:22:04 Alan: why don't we see what happens next week? 18:22:29 Bijan: the issue is that one can hardly expect to get new information to change mind 18:22:41 q? 18:22:58 Alan: it's not about changing minds but about other people getting information to understand what's going on 18:23:26 IanH: let's wait until the next week if additional information appears, if not, just call for a vote 18:23:38 q? 18:23:43 +1 18:23:44 q+ 18:23:46 Subtopic: Issue 138 (name of date datatype) 18:23:48 +1 to owl:datetime 18:23:51 Zakim, mute me 18:23:51 bmotik should now be muted 18:24:15 q? 18:24:19 q? 18:24:20 We are waiting for the response to Peter's email 18:24:22 ack pfps 18:24:33 Peter: perhaps we should put this in some documents 18:24:52 Yes 18:24:58 ... not as a resolved but just to make sure it's not forgotten 18:24:58 q? 18:25:01 There is an editor's note 18:25:08 PFPS: owl:dateTime would be the save choice 18:25:16 s/save/safe/ 18:25:19 ... this would be in syntax, Boris says it would there 18:25:55 Alan: (?) what is the definition of punning at the moment? 18:25:58 q? 18:26:10 I think it's what peter says it was 18:26:11 shouldn't there be an email discussion in the past about the "which punning" question? 18:26:28 ... there are a few definitions going, so which is the one we subscribe to? to explain it to people 18:26:40 IanH: other issues are probably longer to discuss 18:26:42 Topic: AOB 18:26:53 IanH: no additional items, so let's conclude 18:26:54 -msmith 18:26:55 bye 18:26:56 -bmotik 18:26:56 bye bye 18:27:01 -uli 18:27:01 bye 18:27:02 -baojie 18:27:02 -Peter_Patel-Schneider 18:27:04 -bijan 18:27:04 -Sandro 18:27:05 msmith has left #owl 18:27:05 -Achille 18:27:07 thank, Ian :-) 18:27:08 -IanH 18:27:09 -Alan 18:27:10 -m_schnei 18:27:13 -bcuencagrau 18:27:25 And thanks to you too :-) 18:30:31 Sandro, could you help me with getting the minutes on wiki? 18:30:52 got to http://www.w3.org/2008/09/10-owl-irc.txt but can't create a wiki page 18:31:02 what happens when you try? 18:31:34 permissions on editing http://www.w3.org/2007/OWL/wiki/Chatlog_2008-09-10 18:32:07 Who are you logged into the wiki as? (first word on the very top line) 18:32:22 as me (MartinDzbor) 18:32:50 -MartinD 18:33:20 And when you visit http://www.w3.org/2007/OWL/wiki/Chatlog_2008-09-10 is there an "edit" tab? 18:34:04 oi.... wiki timeout :-( 18:34:11 now I got there... 18:39:12 -Zhe 18:39:13 SW_OWL()1:00PM has ended 18:39:14 Attendees were +0190827aaaa, MartinD, Sandro, IanH, bmotik, Zhe, m_schnei, Achille, uli, bcuencagrau, +1.518.276.aacc, baojie, Alan, +1.202.408.aadd, msmith, Peter_Patel-Schneider, 18:39:17 ... bijan 19:00:07 MartinD has left #OWL 20:08:27 alanr_ has joined #owl 20:52:29 Zakim has left #owl 22:35:43 alanr has joined #owl 22:53:53 alanr has joined #owl 23:14:40 sandro has joined #owl