Chatlog 2009-05-06

See original RRSAgent log and preview nicely formatted version.

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

<scribenick> PRESENT: bmotik, IanH, Peter_Patel-Schneider, MarkusK_, bcuencagrau (muted), Rinke (muted), Zhe (muted), Sandro, zimmer, alanr, uli, michael schneider, baojie, evan, mike smith, achille
<scribenick> REGRETS: Elisa, Christine
<scribenick> CHAIR: Ian Horrocks
17:00:52 <IanH> ScribeNick: bmotik
17:01:53 <bmotik> Topic: Agenda amendments
17:02:02 <bmotik> IanH: Any admendments?
17:02:13 <bmotik> Ianh: No amendments
17:02:19 <bmotik> Topic: Previous minutes
17:02:25 <bmotik> Ianh: Anyone looked at that?
17:02:28 <MarkusK_> the minutes are okay, but have some "FIXME"s
17:02:30 <pfps> minutes looked acceptable to me
17:02:31 <bmotik> Ianh: seemed alright to me.
17:02:54 <bmotik> Ianh: Can we accept them but ask Michael to do a bit more of tidying up
17:03:06 <pfps> I think that even the FIXMEs were mostly OK - none of them were related to what I said, however,
17:03:13 <bmotik> Topic: Pending review actions
17:03:25 <MarkusK_> +1 most FIXMEs could just be deleted
17:03:34 <bmotik> Ianh: ACTION-333 seems to be completed
17:03:50 <bmotik> IanH: ACTION-335 is on Bijan, but he is not on the call
17:04:11 <bmotik> Ianh: Let's push it off to next week
17:04:17 <bmotik> Topic: Documents and reviewing
17:04:32 <bmotik> Ianh: I see that the comment to rdf:text has been agreed to and sent
17:04:39 <pfps> +1 to say that the response wasn't completely positive
17:05:58 <pfps> +1 to let Boris fight this out with the commenter
17:06:00 <sandro> boris: michael sperberg-mcqueen seemed okay with all of our comment except one part, where he seems to disagree.   I've asked him on the rdf-text list if it's okay for us to just explain better the scope of what we're doing.    I've also asked him for more information on what rdf:text doesn't handle. 
17:06:42 <alanr> ruby is annotation
17:06:47 <IanH> sandro: e.g., directions of languages (c.f. Ruby)
17:06:56 <alanr> didn't seem to me to be part of the utterance
17:07:15 <bmotik> Sandro: For example, in Japanese, one might need additional markup for direction and so on
17:07:22 <bmotik> Sandro: I agree that we should not handle that
17:07:39 <pfps> saying that rdf:text is just about language tagging sounds like the right solution to me
17:07:47 <bmotik> Sandro: I am hoping that someone from the i18n community will be able to provide a response
17:08:06 <bmotik> Ianh: So Boris is going to continue handling this with Axel?
17:08:12 <bmotik> bmotik: yes
17:08:57 <bmotik> alanr: We could say a bit more about the Ruby text and the direction. The distinction seems to be about layout and annotation. I could draft something that explains this.
17:10:25 <bmotik> bmotik: If you can provide some text for the introduction about what we handle and what we don't handle, that would be appreciated. We should not start speculating as to how one might want to handle these problems.
17:09:28 <alanr> +1, document shouldn't say. Response should
17:09:39 <alanr> will do
17:10:08 <pfps> the suggested text should be produced ASAP
17:10:08 <alanr> action alanr to draft some text for rdf:text or response re unhandled features mentioned by sperber 
17:10:26 <alanr> understood, peter
17:10:35 <bmotik> ianh: Other comments: there was one with a typo.
17:10:42 <bmotik> ianh: This was fixed.
17:10:59 <bmotik> ianh: There was a comment about the use of the word "axiom" in the Primer.
17:11:29 <MarkusK_> no, the response does not entail changes to the Syntax
17:11:34 <bmotik> bmotik: Does this entail any change to the Syntax?
17:11:42 <bmotik> ianh: No, it is only in the Primer.
17:11:48 <MarkusK_> we stick to "axiom" but avoid some unfortunate formulations about "deriving axioms"
17:12:21 <MarkusK_> peter: I looked at the response draft and I think it is good as it is
17:12:37 <bmotik> IanH: I have checked with Richard about the comment of "Class" as is used in the documents, and he does want to consider his e-mail to be a LC comment. We need to produce a response, probably about the lines of "that's standard in our world".
17:13:00 <alanr> sandro: could you please fix my action item syntax?
17:13:07 <bmotik> Topic: Quick Reference Guide
17:13:17 <bmotik> ianh: There has been quite a lot of activity on QRG
17:13:34 <bmotik> ianh: Jie implemented the resolution regarding the three-column format.
17:13:49 <bmotik> ianh: There has been discussion about other aspects of the document.
17:13:59 <bmotik> ianh: Can the people who are working on the document fill us in?
17:14:09 <bmotik> ianh: Jie, what's left to figure out?
17:14:15 <baojie> http://www.w3.org/2007/OWL/wiki/Talk:Quick_Reference_Guide#May_5.2C_2009
17:14:19 <bmotik> baojie: There are a couple of pending issues/
17:14:39 <bmotik> baojie: The first thing is how do we list OWL 2 Full vocabulary in the reference.
17:15:01 <baojie> http://www.w3.org/2007/OWL/wiki/Talk:Quick_Reference_Guide#Additional_Vocabulary_in_OWL_2_Full
17:15:03 <bmotik> baojie: I have an initial proposal that we have an additional table.
17:15:18 <schneid> Jie talks about vocabulary terms, that are exclusively used in the RDF-Based Semantics
17:15:23 <bmotik> baojie: This table would list all vocabulary that is not used in the syntax.
17:15:50 <bmotik> ianh: I looked at the link but didn't understand it.
17:15:53 <pfps> Section name is fine, as long as it is in the Appendix section.
17:16:03 <bmotik> baojie: The additional vocabulary in OWL 2 Full
17:16:13 <bmotik> ianh: Which of the ones do you think are used in OWL 2 Full?
17:16:18 <bmotik> baojie: The second column.
17:16:27 <schneid>  /all/ of these terms occur in the reverse Mapping, I believe!
17:16:40 <bmotik> baojie: I find that there is a mismatch in the syntax and and the RDF semantics.
17:16:58 <pfps> They have to, as they show up in OWL 1 ontologies
17:17:20 <bmotik> baojie: The second thing is that there is an issue of naming: shall we call things "onotlogy properties" or "annotation properties">
17:17:21 <schneid> call them annotation properties, and don't talk about owl:OntologyProperty at all in the QRG
17:17:31 <schneid> it will only confuse people
17:17:34 <Rinke> +1 to michael
17:17:38 <bmotik> baojie: Peter suggested to remove n-ary features.
17:17:46 <bmotik> baojie: I don't see a strong reason to do that.
17:18:06 <bmotik> baojie: Peter had some other comments that I haven't had the chance to look at.
17:18:27 <bmotik> pfps: Fourth thing: Uli suggested a reorganization of the sections.
17:18:16 <uli> done
17:18:33 <bmotik> baojie: I've already done that.
17:18:33 <uli> that was very speedy indeed!
17:19:01 <baojie> OWL 2 Full Vocabulary in QRG
17:19:23 <bmotik> ianh: So what do we want to do about OWL 2 vocabulary?
17:19:23 <pfps> also organization of 2.7 and 2.8 (Annotations and Ontologies)
17:20:06 <bmotik> schneid: Some of this vocabulary should go to a separate table.
17:20:21 <bmotik> schneid: Ignore ontology properties in QRG and call them annotation properties as in the Syntax.
17:20:56 <bmotik> schneid: For me, this is the RDF semantics feature.
17:21:38 <bmotik> ianh: Michael seems to believe that we should just leave this vocabulary out. Jie, are you OK with that?
17:21:35 <pfps> I'm all *for* it
17:21:52 <bmotik> baojie: I'm OK with that.
17:22:04 <bmotik> ianh: So then we've decided what to do about Item 1
17:22:08 <schneid> schneid: there are two groups: the group of terms that are mapped by the reverse RDF mapping (DataRange, distinctMembers, DeprecatedClass|Property)
17:22:27 <bmotik> baojie: Can we specify what is *this* vocabulary?
17:22:29 <schneid> schneid: the other "group" is the term owl:OntologyProperty
17:22:49 <bmotik> ianh: If I understood Micahel, I presume that he thinks the OWL 2 vocabulary should not be listed in the QRG
17:23:01 <bmotik> pfps: Precisely that
17:23:14 <baojie> DataRange, distinctMembers, DeprecatedClass|Property owl:OntologyProperty 
17:23:41 <bmotik> schneid: Don't talk about OWL 2 ontology properties; just call them annotation properties
17:23:42 <schneid> schneid: (1) don't talk about owl:OntologyProperty, (2) talk about the actual ontology properties as annotation properties, because they are called so in all other documents except the RDF-based Semantics
17:24:02 <bmotik> baojie: Peter suggested to move the ontology properties into the annotations section
17:24:11 <bmotik> schneid: Yes, do that
17:24:16 <bmotik> Ianh: All clear?
17:24:20 <bmotik> baojie: Yes
17:24:39 <bmotik> ianh: About the third point (n-ary)... opinions?
17:24:40 <pfps> -0.1
17:24:47 <bmotik> -0.5
17:25:05 <alanr> -1
17:24:57 <bmotik> ianh: It seems unnecessary because they are not a part of the language
17:25:26 <ewallace> since it's not a "real" feature of OWL2 it seems a good sacrifice for shortening QRG
17:25:26 <msmith> unnecessary since they can't be used
17:25:32 <pfps> n-ary some and all are in the syntax, but are not useful unless you have the DRE
17:25:35 <bmotik> baojie: But they do add something to the syntax, so if we leave them out, this is the only construct that we leave out
17:25:46 <bmotik> ianh: But in Syntax, you are allowed to have only one property in Syntax
17:25:55 <bmotik> pfps: We are talking about n-ary some and all
17:26:26 <bmotik> pfps: They are in the Syntax, but they are not useful if you don't have the n-ary extension. And this extension is not a part of the OWL 2 language.
17:26:31 <msmith> an nary extension that might have its own QRG
17:26:58 <bmotik> pfps: Nobody should be writing this unless they also take into account the n-ary extension.
17:27:03 <ewallace> So this is not "Quick Reference" material
17:27:30 <pfps> we are talking about DataAllValuesFrom(:age :height drs:greater)
17:27:31 <bmotik> baojie: So we could not have it in the expression table, but we should have it in the new features list
17:27:37 <bmotik> baojie: It is mentioned in NF&R
17:27:56 <pfps> a pointer to NF&R is OK
17:28:02 <bmotik> IanH: So everybody seems to agree that we shouldn't have it in the syntax table, OK?
17:28:04 <bmotik> baojie: OK
17:28:14 <bmotik> ianh: We should see about the new features table
17:28:33 <msmith> yes is not OWL 2 DL
17:28:36 <bmotik> schneid: A document that uses an n-ary datatype is not a conformant OWL 2 ontology document
17:29:11 <bmotik> schneid: When we use n-ary properties, we need to use an n-ary datatype; this datatype is not in the list in the Syntax; ergo, the containing document is not an OWL 2 document.
17:30:05 <bmotik> schneid: Implementors should not support these terms and should still create conformant OWL 2 reasoners.
17:30:45 <bmotik> ianh: The question left open is whether we should mention n-ary in the "new features" section. It seems to me that we shouldn't.
17:30:57 <bmotik> baojie: I'd like to be consistent with the NF&R document.
17:31:28 <Rinke> I think the NF&R doc needs to be changed...
17:31:31 <bmotik> baojie: If this feature is in NF&R, then it should be in QRG, and vice versa.
17:32:04 <pfps> I'm happy leaving the entry under 4.1 class expressions on n-ary.  I think that the entry under 4.1 / data ranges should go.
17:32:16 <alanr> seems inconsistent to have one and not the other.
17:32:40 <alanr> prefer not, but can live with it if there are strong opinions
17:32:43 <pfps> They both point to the same place - but they both *can* stay
17:33:02 <pfps> QRG is now *much* better!
17:33:09 <baojie> thanks!
17:33:49 <pfps> Work might be needed on Annotations and Ontologies
17:33:50 <bmotik> pfps: Jie and I can argue the changes regarding the annotations between us.
17:33:51 <bmotik> Topic: More issues with rdf:text
17:33:51 <bmotik> We have got a response from Philipps Addison regarding rdf:text. Can we discuss this again?
17:34:27 <uli> Boris, where is the response?
17:34:58 <ewallace> Is this about rdf:statements?
17:34:59 <IanH> http://lists.w3.org/Archives/Public/public-owl-comments/2009May/0007.html
17:35:49 <uli> Boris, does he understand the difference between 'internationalised text' and 'internationalised utterances'?
17:35:54 <bmotik> bmotik: The response by Phlips Addison is not as we want it.
17:36:06 <alanr> I have sent analysis of ruby text and bidi issue to list, dispatching my earlier action
17:36:13 <bmotik> ACTION: sandro to Look at the response for rdf:text
17:36:13 <trackbot> Created ACTION-336 - Look at the response for rdf:text [on Sandro Hawke - due 2009-05-13].
17:36:34 <bmotik> bmotik: So I'm off the hook for this?
17:36:40 <bmotik> ianh: For the moment...
17:36:44 <bmotik> Topic: NF&R
17:36:52 <bmotik> ianh: Christine is not on the call today
17:36:59 <bmotik> ianh: I saw a detailed review from Rinke
17:37:12 <bmotik> pfps: I think the document is getting close.
17:37:12 <pfps> I think that NF&R is getting close.
17:37:31 <bmotik> rinke: My suggestions are editorial-ish, but might be controversial.
17:37:41 <ewallace> Christine finds Rinke's comments controversial.
17:38:07 <bmotik> pfps: I'm not sure I agree with Rinke's probles of appendix vs. non-appendix.
17:38:17 <bmotik> pfps: Maybe I'll respond to Rinke's message.
17:38:27 <bmotik> ianh: I'd like to minimize changes if possible.
17:38:41 <bmotik> rinke: I don't advocate a large change.
17:39:04 <bmotik> ianh: So Peter and Rinke and Christine are dealing with that?
17:39:07 <bmotik> pfps: Yes
17:39:10 <MarkusK_> Primer is ready for review
17:39:17 <bmotik> Topic: Primer
17:39:26 <bmotik> ianh: It is getting close to reviewing.
17:39:19 <alanr> Who are the reviewers?
17:39:26 <msmith> I am one of the reviewers
17:39:31 <bmotik> ianh: Mike volunteered to review it.
17:39:44 <bmotik> ianh: Michelle offered to review it and Deborrah.
17:39:57 <bmotik> ianh: I don't think there is lots more to say aobut it.
17:39:58 <MarkusK_> +1 Primer is ready, nothing more to be said
17:40:11 <bmotik> Topic: Overview and Manchester Syntax
17:40:17 <bmotik> ianh: They are ready a while ago
17:40:34 <bmotik> ianh: They were held bak because of keeping the UFD back because of the schedule
17:40:41 <bmotik> Topic: Data-range Extension
17:40:44 <uli> yes
17:40:49 <bmotik> ianh: It might need further work
17:41:08 <bmotik> Ianh: Uli, ETA?
17:41:15 <bmotik> Uli: ETA?
17:41:24 <bmotik> Uli: There is not much to do on it.
17:41:49 <bmotik> Uli: I could say "next week", but then it might not be the case
17:42:04 <bmotik> ianh: Since it is a note only, it will not impact on our schedule
17:42:09 <bmotik> Topic: References in the documents
17:42:21 <bmotik> ianh: Sandro has some magic method that we used on the Overview
17:42:28 <bmotik> pfps: I don't like Sandro's solution
17:42:38 <bmotik> pfps: It doesn't show up in the Wiki
17:42:50 <bmotik> pfps: I like Michael's solution, which is based on templates.
17:42:58 <MarkusK_> +1 to use templates for references
17:43:18 <bmotik> ianh: Yes, I used this in Comformance
17:43:44 <pfps> see http://www.w3.org/2007/OWL/wiki/References
17:43:46 <ewallace> Reference template solution sounds cool.
17:44:04 <bmotik> ianh: Sandro, do you want to defend your solution?
17:44:12 <bmotik> ianh: What is your solution actually?
17:44:34 <bmotik> Sandro: You can do all that at publication time
17:45:02 <ewallace> Fix once, correct everywhere!
17:45:04 <bmotik> Sandro: You can do it anyway you want as long as you get it right for the publication time
17:45:23 <bmotik> ianh: The point is to ensure consistency among references
17:45:33 <schneid> true, there is a lot of manual work in the documents to always write the correct citation marks!
17:45:58 <bmotik> Sandro: We haven't decided what the text in square brackets should be, and I tried to automate that
17:46:09 <bmotik> pfps: We could have another template for the pointer.
17:46:10 <schneid> yes, peter, this was my idea too
17:46:16 <bmotik> pfps: I don't think I want to go there.
17:46:28 <bmotik> pfps: Of course, what we need is LaTeX
17:46:39 <bmotik> ianh: But we don't have that
17:46:44 <alanr> blahdeblahde will help with wiki issues
17:47:00 <alanr> meaning *I* will help with them if I'm made aware of issue
17:47:01 <schneid> there should always be a /pair/ of templates for each reference
17:47:18 <MarkusK_> Boris: My impression is that templates slow down page creation.
17:47:32 <pfps> What should the citations in the text really look like?  RIght now we generally use "FooLong [FooShort]" or something like that.
17:47:33 <bmotik> bmotik: The Wiki is slow and this is because of the many templates
17:47:33 <MarkusK_> -1 to Boris: pages are cahced unless you edit them
17:47:49 <bmotik> Sandro: The pages should be cached
17:47:54 <pfps> Wiki is just slow
17:48:08 <bmotik> bmotik: Then Wiki is just a *bad* tool
17:48:42 <sandro> (the load average on the wiki server is currently: top - 17:48:33 up 235 days, 22:54,  1 user,  load average: 12.88, 6.58, 3.78 )    :-(   :-(   :-(
17:49:12 <bmotik> ianh: So we are using templates
17:49:25 <bmotik> Sandro: It is the Wiki server that is overloaded; it is not the technology
17:49:48 <bmotik> Topic: Implementation & Testing
17:49:51 <pfps> Editors should (sometime) try to templatize their references!
17:49:55 <bmotik> ianh: Anything interesting happening?
17:50:03 <MarkusK_> I have nothing interesting to report on tests
17:50:20 <msmith> difference since last week in test suite http://www.w3.org/2007/OWL/wiki/index.php?title=Test_Suite_Status&diff=23288&oldid=22789
17:50:30 <bmotik> alanr: I sent a note out about some syntax tests that I have up and running. I've proposed a suggestion as to how a reasoner might handle them.
17:51:05 <bmotik> alanr: I suggest we check these tests in two ways. (1) We check that the tripes are isomorpic. (2) We do a bidirectional entailment test.
17:51:24 <bmotik> alanr: I am in the process of figuring out how to transfer these tests into the WIki.
17:51:26 <pfps> http://lists.w3.org/Archives/Public/public-owl-wg/2009Apr/0440.html ???
17:51:38 <alanr> peter, yes
17:51:43 <bmotik> alanr: We don't have a category for syntax tests. I need to talk about this to Mike.
17:52:04 <bmotik> schneid: I've started the work on test cases. Currently, evertyhing is in the FZI SVN.
17:52:18 <bmotik> schneid: The RDF semantics will be covered.
17:52:34 <bmotik> ianh: This will involve you pouring at some stange a large number of tests, right?
17:52:52 <bmotik> schneid: The best I can do is produce the page code and then put it into the Wiki.
17:53:01 <bmotik> schneid: I would like to see bulk-upload if possible.
17:53:29 <bmotik> ianh: So you don't need any structural change of the tests; it is just the upload question?
17:53:47 <bmotik> schneid: There are some bugs.
17:54:02 <bmotik> schneid: Other than that, I'm fine.
17:54:36 <pfps> The pointer to svn.mumble.net doesn't appear to be working
17:55:03 <alanr> just clicked on it successfully. FIrewall issue?
17:55:12 <alanr> http://svn.mumble.net:8080/svn/lsw/trunk/owl/owl2/tests/fs2rdf/
17:55:13 <bmotik> msmith: There are 182 tests with no status. The vast majority of those are incomplete. E.g., these tests miss the species, the profile, or something.
17:55:24 <schneid> Just to be clear, I will create the final test case format automatically, so they should be fine structurally
17:55:27 <bmotik> msmith: I'll try to pour over these next week and ask people to finish tests.
17:55:45 <bmotik> msmith: This is a bunch of tests for a few people.
17:55:51 <schneid> schneid: I suggest to have a "OPTIONAL" marker
17:56:07 <MarkusK_> schneid: feel free to contact me regarding possible bulk upload
17:56:17 <schneid> ok, markus
17:56:20 <bmotik> ianh: The two Michaels should have a discussion about the "OPTIONAL" feature
17:56:23 <MarkusK_> ... there are options, but I need to knwo what is convenient to you
17:56:28 <bmotik> IanH: Coming back to the syntactic tests...
17:56:41 <bmotik> ianh: Alan, does syntactic tests require some change to the structure of the test harness?
17:56:42 <bmotik> alanr: (Explains the differences which are not scribed)
17:57:37 <uli> I am not sure I understood
17:57:45 <pfps> I also did not understand what was wanted.
17:57:52 <alanr> test categories: give functional, rdf expect correct
17:57:56 <uli> Alan, are those test like the ones you described in earlier emails?
17:57:59 <alanr> give function, rdf expect incorrect
17:58:09 <alanr> give rdf, expect rejection (syntax error)
17:58:25 <schneid> schneid: I am in the process of writing a suite of test cases for the RDF-Based Semantics, but currently everything is locally stored at FZI
17:58:50 <pfps> Are these tests?  I don't understand what is supposed to be going on.
17:58:57 <schneid> schneid: I have just look through the TestCases section of Conformance, and found a few issues, will report them soon
17:59:00 <bmotik> alanr: THe tests in the earlier e-mails were only where you specify functional and you want to get a correct translation. The new tests also contain syntax errors that a parser should detect.
17:59:25 <bmotik> msmith: What is the difference between an RDF graph that a parser rejects as incorrect or an RDF graph that a parser says "It is not an OWL 2 DL ontology"?
17:59:51 <bmotik> alanr: If we have at the end of the RDF mapping we have excess triples -- that's the thing I want to test for.
18:00:04 <bmotik> msmith: But that graph is also not an OWL 2 DL ontology.
18:00:22 <pfps> I want to see a proposal for what needs to be changed.
18:00:24 <bmotik> alanr: It's a different case because you can not be an OWL 2 DL ontology because you've violated something else.
18:00:42 <bmotik> ianh: I think we should hold off further discussion until we see some tests from Alan.
18:00:47 <bmotik> pfps: I didn't see any examples.
18:01:22 <msmith> alanr's previous link did work for me http://svn.mumble.net:8080/svn/lsw/trunk/owl/owl2/tests/fs2rdf/
18:02:05 <bmotik> pfps: I'm happy to wait until I see some proposal as to what needs to be changed in the test set up.
18:02:07 <pfps> still not working for me
18:02:23 <bmotik> Topic: Features at risk
18:02:29 <bmotik> ianh: That's just a standard reminder
18:02:38 <bmotik> Topic: Any other business
# SPECIAL MARKER FOR CHATSYNC.  DO NOT EDIT THIS LINE OR BELOW.  SRCLINESUSED=00000593