IRC log of owl on 2008-10-23

Timestamps are in UTC.

06:55:50 [RRSAgent]
RRSAgent has joined #owl
06:55:50 [RRSAgent]
logging to
06:56:07 [pfps]
Zakim, this will be owlwg
06:56:07 [Zakim]
ok, pfps; I see SW_OWL(F2F)2:30AM scheduled to start 26 minutes ago
06:56:21 [pfps]
RRSAgent, make records public
06:56:33 [pfps]
ScribeNick: pfps
06:56:36 [Elisa]
Elisa has joined #owl
06:56:59 [pfps]
06:58:12 [Zakim]
SW_OWL(F2F)2:30AM has now started
06:58:19 [Zakim]
06:58:25 [Elisa]
07:04:19 [IanH]
IanH has joined #owl
07:06:56 [bcueencagrau]
bcueencagrau has joined #owl
07:07:07 [sandro]
sandro has joined #owl
07:07:35 [MarkusK_]
MarkusK_ has joined #owl
07:07:40 [pfps]
IanH: Welcome (to ...)
07:07:51 [sandro]
zakim, who is here?
07:07:51 [Zakim]
On the phone I see Elisa_Kendall
07:07:52 [Zakim]
On IRC I see MarkusK_, sandro, bcueencagrau, IanH, Elisa, RRSAgent, Zakim, pfps, ivan, trackbot
07:08:37 [wallace]
wallace has joined #owl
07:10:20 [sandro]
PRESENT: Ian, Boris, Peter, Bernardo, Sandro, MarkusK, m_schnei, Bijan, Evan, Ivan
07:12:45 [m_schnei]
m_schnei has joined #owl
07:12:49 [sandro]
zakim, call Riviera_B
07:12:49 [Zakim]
ok, sandro; the call is being made
07:12:51 [Zakim]
07:14:09 [dom]
dom has joined #owl
07:14:09 [pfps]
Sandro: nothing on local arrangement
07:14:30 [pfps]
IanH: introductions
07:14:40 [pfps]
...: Hi, I'm me
07:14:42 [ivan]
zakim, dial Riviera_B
07:14:42 [Zakim]
ok, ivan; the call is being made
07:14:43 [Zakim]
07:15:03 [dom]
dom has left #owl
07:15:31 [bmotik]
bmotik has joined #owl
07:15:51 [pfps]
Observers - Henson Graves, Jeremy Carroll
07:16:09 [sandro]
Observers: Henson_Graves, Jeremy_Carroll
07:16:41 [sandro]
07:16:42 [pfps]
IanH: Timeline (follow link in agenda)
07:17:13 [pfps]
IanH: I put a real timeine (not T0+)
07:17:38 [pfps]
IanH: We are about 2 months behind the scheduled time for Last Call
07:17:53 [pfps]
IanH: It thus would be good to move forward with due haste
07:18:43 [pfps]
Bijan: The schedule was designed to be aggressive (but with a bit of slack)
07:18:56 [pfps]
Sandro: The slack is ... two months
07:19:09 [pfps]
Topic: Document Status
07:19:19 [pfps]
Polycom: Beep
07:19:34 [pfps]
IanH: This section is intended as a review
07:19:45 [pfps]
IanH: Can the editors say the status
07:19:45 [Zakim]
07:19:49 [sandro]
zakim, who is on the call?
07:19:49 [Zakim]
On the phone I see Elisa_Kendall, Riviera_B
07:20:23 [pfps]
Boris: Syntax is up to date - there are some issues that will impact it
07:20:42 [pfps]
IanH: There has been internal review (for last PWD)
07:20:47 [IanH]
07:20:55 [pfps]
Ivan: There has been no major external comments
07:21:03 [IanH]
zakim, who is here?
07:21:03 [Zakim]
On the phone I see Elisa_Kendall, Riviera_B
07:21:04 [Zakim]
On IRC I see bmotik, m_schnei, wallace, MarkusK_, sandro, bcueencagrau, IanH, Elisa, RRSAgent, Zakim, pfps, ivan, trackbot
07:21:09 [pfps]
Bijan: Does current syntax document meet Evan's needs
07:21:29 [pfps]
Evan: Way better than it was - usable - not great because of organization
07:22:00 [pfps]
Evan: Reorganization is currently for the spec/implementation, not users
07:22:15 [pfps]
Bijan: Not explicitly - we did have discussions on the order
07:22:29 [pfps]
07:22:46 [pfps]
Bijan: There are various organizations of reference docs in the literature
07:22:59 [pfps]
Boris: Currently Syntax is a *reference* document
07:23:00 [Zhe]
Zhe has joined #owl
07:23:34 [pfps]
Evan: Quick Reference Guide could be used as an index
07:23:47 [IanH]
07:24:04 [pfps]
Bijan: Primer can serve as another "index"
07:24:42 [pfps]
Bijan: Three "indexes" - ToC, QRG, Primer
07:24:50 [Zakim]
+ +1.978.692.aaaa
07:24:57 [IanH]
07:25:05 [pfps]
Evan: The problem is using it as a reference
07:25:13 [Zhe]
zakim, +1.978.692.aaaa is me
07:25:13 [pfps]
Evan: The ordering is wrong
07:25:13 [Zakim]
+Zhe; got it
07:25:19 [pfps]
Boris: What is needed?
07:25:20 [IanH]
Hello Zhe!
07:25:25 [Zhe]
zakim, mute me
07:25:25 [Zakim]
Zhe should now be muted
07:25:28 [Zhe]
Hi Ian!
07:25:41 [IanH]
07:25:48 [pfps]
Evan: Things related to object properties grouped together
07:26:11 [pfps]
Boris: But what about domain axioms - they are related to both classes and object properties
07:27:41 [IanH]
07:27:41 [pfps]
Christine: What is under discussion now?
07:28:09 [IanH]
07:28:11 [pfps]
Bijan: We are now discussing the Syntax document - but are also pulling in relationships to other documents
07:28:47 [pfps]
Bijan: Old reference has informal discussions, which are not in the QRG
07:28:59 [sandro]
PRESENT: Ian, Boris, Peter, Bernardo, Sandro, MarkusK, m_schnei, Bijan, Evan, Christine, Rinke, Ivan
07:29:04 [IanH]
07:29:11 [pfps]
IanH: A complete redesign of Syntax is a major effort
07:29:12 [bmotik]
bmotik has joined #owl
07:29:39 [pfps]
Evan: A complete redesign is not in the cards
07:30:00 [sandro]
Evan: I'm fine with using the Quick Reference Guide as the index to Syntax.
07:30:05 [IanH]
07:30:06 [pfps]
Evan: A reference index is needed - either QRG or part of the document
07:30:40 [pfps]
IanH: OK, syntax is in pretty good shape, modulo outstanding issues and perhaps an index
07:30:49 [pfps]
SubTopic: Semantics Document
07:30:59 [IanH]
07:31:13 [pfps]
Boris: Similar status to syntax - up to date - outstanding issues may need changes
07:31:35 [bijan]
bijan has joined #owl
07:31:38 [sandro]
pfps: It's our contention that the Direct Semantics current correctly describes the meaning of OWL.
07:32:02 [sandro]
ian; Finished, modulo outstanding issues.
07:32:05 [pfps]
SubTopic: RDF Semantics
07:32:06 [IanH]
07:32:15 [pfps]
Michael: RDF Semantics is a bit behind
07:32:18 [sandro]
07:32:30 [pfps]
Michael: There are a couple of minor things that need to be added
07:32:43 [pfps]
Michael: The two documents are structurally aligned
07:33:31 [IanH]
07:33:44 [pfps]
Michael: Outstanding issues - correspondence theorem, test cases that exercise rdf semantics
07:33:51 [sandro]
m_schnei: the correspondence theorem proof still needs work.
07:34:10 [pfps]
Bijan: Do we believe that the theorem is correct - if so then we should be able to go to last call - if not then we need to worry
07:34:19 [IanH]
07:34:32 [pfps]
Michael: I believe the theorem and that it is a good as we can get
07:34:55 [sandro]
bijan: Are any proof errors such that the language would have to change?
07:35:00 [IanH]
07:35:05 [sandro]
m_schnei: I don't think the language will have to change.
07:35:28 [pfps]
Bijan: Do you think that the semantics is OK
07:35:29 [Rinke]
Rinke has joined #owl
07:35:35 [pfps]
Michael: 95 per cent
07:35:56 [pfps]
Subtopic: Conformance and Test Cases
07:36:07 [pfps]
IanH: This could be more contentious
07:36:15 [pfps]
Ivan: We need more test cases
07:36:34 [IanH]
07:36:42 [pfps]
IanH: Mike Smith wants to participate
07:37:00 [pfps]
Bijan: When I wanted to submit test cases the structure wasn't redy
07:37:12 [pfps]
Ivan: What is the experience of the OWL 1 test cases
07:37:47 [pfps]
Bijan: They are great, much better than before, they help a lot in checking initial part of implementation
07:38:13 [IanH]
07:38:16 [pfps]
Markus: Most OWL 1 test cases have been copied over
07:38:35 [pfps]
Ivan: We might only need tests for the new features
07:38:49 [pfps]
Bijan: We could do more, but getting to the OWL 1 level is adequate
07:39:17 [pfps]
IanH: There was also fitting into Lite, DL, Full, so the tests need to be remarked
07:39:33 [IanH]
07:39:44 [pfps]
IanH: We also probably need test cases to check the boundaries of the profiles
07:39:46 [MarkusK_]
Test cases
07:40:08 [MarkusK_]
This page contains links to lists showing all test cases, by various criteria
07:40:59 [pfps]
Pfps: What about the status of T&C itself
07:41:09 [pfps]
Bijan: We will ask for tests at OWLED
07:41:22 [pfps]
Ivan: We need test cases ready for CR
07:41:41 [pfps]
Bijan: Not so - test cases could come out of CR - we need a reasonable set going in
07:41:45 [IanH]
07:42:23 [pfps]
Jeremy: OWL 1 test cases lagged going into LC by two months
07:42:33 [IanH]
07:42:55 [pfps]
Sandro: At some time there has to be a set of approved test cases
07:43:10 [pfps]
IanH: Mike Smith wants a process for approving new test cases
07:43:27 [pfps]
Sandro: Initially by hand, then we can use implementations to help approval process
07:43:51 [pfps]
Jeremy: OWL 1 document included the process for approving test cases
07:44:20 [pfps]
IanH: Conformance part has been approved - and has no outstanding issues
07:44:32 [pfps]
Subtopic: RDF Mapping
07:44:39 [pfps]
IanH: What about RDF Mapping?
07:45:10 [pfps]
Boris: Same status as Syntax and Semantics - up to date - some outstanding issues
07:45:23 [IanH]
07:46:13 [pfps]
Ivan: Looking at the QRG there appear to be some mismatches between functional and RDF syntaxes
07:47:01 [IanH]
07:47:27 [pfps]
Boris: There are reasons for some of the mismatches
07:47:36 [IanH]
07:47:44 [pfps]
ACTION: pfps to check differences between functional and RDF syntaxes
07:47:44 [trackbot]
Created ACTION-232 - Check differences between functional and RDF syntaxes [on Peter Patel-Schneider - due 2008-10-30].
07:48:12 [IanH]
07:48:19 [pfps]
Ivan: also XML syntax
07:48:54 [pfps]
Bijan: XML syntax mirrors functional syntax
07:49:04 [pfps]
Subtopic: XML serialization
07:49:17 [schneid]
schneid has joined #owl
07:50:13 [IanH]
07:50:27 [pfps]
Bijan: Document is up to date - potential outstanding issues
07:50:49 [pfps]
Bijan: Would be nice to have a non-normative RelaxNG syntax
07:51:10 [IanH]
07:51:26 [pfps]
Bijan: This would an editorial addition - non critical - could even be after last call
07:51:49 [IanH]
07:52:20 [pfps]
Bijan: Issues with aspects of design - too verbose - need to check with Matt Horridge
07:52:42 [pfps]
IanH: There was a query from Alan related to the MOF metamodel - can we generate the syntax from the MOF?
07:53:25 [Elisa]
What would be generated from the MOF metamodel is XMI, which is an OMG specification for XML schema interchange
07:53:32 [pfps]
Bijan: Not a good idea - no evidence that it would work - know that conversion to RelaxNG works
07:53:46 [pfps]
Bijan: MOF conversion to XML might result in an unreadable schema
07:53:58 [pfps]
Boris: Could end up very close
07:54:06 [Elisa]
This could be mapped to various other surface syntaxes in an automated way
07:54:28 [IanH]
07:54:41 [pfps]
Bijan: I want to see the output before I determine whether it is a good idea
07:55:02 [sandro]
Sandro: We can just wait until someone comes forward wanting this, and see if they're offering to do it.
07:55:24 [pfps]
Elisa: Lots of tools generate XML Schema from a metamodel - could be verbose
07:55:32 [bijan]
07:55:36 [sandro]
elisa: The XMI -- the automatic XML schema -- will be generated automatically by any decent UML tool -- but the XMI has extra cruft, which you'd have to map out of it.
07:55:51 [IanH]
07:55:56 [pfps]
Elisa: What does the WG want to do with the result?
07:56:08 [bijan]
07:56:35 [pfps]
Boris: Why do we want XMI? We then get an automatically-generated syntax
07:56:47 [pfps]
Boris: Depends on result of metamodel issue
07:57:17 [sandro]
Bijan: The question is whether this would result in a better schema. More accurate, ...?
07:57:21 [pfps]
Bijan: I can see point related to above claim. However, is the result a better schema?
07:57:52 [pfps]
Bijan: I would prefer RelaxNG but I'm not proposing to change at this point.
07:57:59 [pfps]
Bijan: We need to be sure of the benefit.
07:58:24 [pfps]
Boris: Peter Hasse sent me an automatically generated schema - it wasn't pretty.
07:58:52 [pfps]
Boris: Peter Hasse said that the generation can be controlled, so maybe a good schema could result
07:59:15 [pfps]
Boris: In any case this depends on the metamodel issue and then a benefits analysis
07:59:21 [pfps]
Bijan: Agree
07:59:35 [pfps]
IanH: Agree and also worry about timeline
07:59:45 [pfps]
Bijan: Can we test whether our schema matches the metamodel
08:00:09 [pfps]
Elisa: Yes, but I'm not up on the tools - I do know someone who knows how to do this
08:00:26 [pfps]
Bijan: Testing our Schema would be a good idea
08:00:47 [IanH]
08:01:16 [pfps]
Elisa: This can also be a debugging tool
08:01:52 [pfps]
Elisa: ECLIPSE has tools that help working on ontologies
08:03:19 [pfps]
Evan: The tools check XMI not Schema
08:04:06 [pfps]
Bijan: But tools turn metamodels into XML Schema - what about doing the reverse?
08:04:31 [pfps]
Evan: The tools result in ugly schema
08:04:41 [pfps]
Bijan: So there are no recognizers?
08:04:55 [sandro]
wikipedia says "exchanging files between UML modeling tools using XMI is rarely possible."
08:05:18 [IanH]
08:05:40 [pfps]
Boris: If we can automatically generate a nice Schema from the metamodel then we get automatic correspondence
08:05:53 [Rinke]
(sandro, that's my personal experience as well)
08:06:25 [IanH]
08:06:36 [pfps]
Bijan: Correctness (consistence) is the only benefit, I believe the schema over the metamodel
08:07:17 [pfps]
Sandro: If what Boris is saying works, then we get some increment to confidence
08:07:25 [pfps]
IanH: Not critical path
08:08:03 [pfps]
Rinke: There are tools that generate metamodel from XML Schema
08:08:07 [sandro]
Sandro: If we can generate a schema from the metamodel, then run it against all the test cases, that would be a nice validation of the metamodel.
08:08:16 [pfps]
IanH: Also a good idea, but not on our critical path
08:08:27 [pfps]
Evan: What is the canonical form of an OWL 2 ontology
08:08:51 [pfps]
Boris: The metamodel (but this is not completely formally defined)
08:09:33 [pfps]
Subtopic: Profiles
08:09:33 [sandro]
Boris: the metamodel -- in natural language, UML, functional syntax etc -- spread through all these bits -- that's the metamodel, and it's the canonical form.
08:09:47 [pfps]
Boris: Up to date - some outstanding issues
08:09:57 [IanH]
08:10:11 [pfps]
Bijan: What about descriptive stuff on the various profiles?
08:10:25 [pfps]
IanH: I added some of this stuff - it is controversial
08:10:52 [sandro]
PRESENT: Ian, Boris, Peter, Bernardo, Sandro, MarkusK, m_schnei, Achille, Bijan, Evan, Christine, Rinke, Ivan
08:10:54 [pfps]
Ivan: Want full grammars for each profile
08:10:57 [bijan]
Editor's Note: This appendix will contain the full grammars of each of the profiles. The grammar will be completed when the technical work on each of the profiles has been finished.
08:11:01 [IanH]
08:11:10 [bijan]
08:11:17 [pfps]
Boris: Editorial note - will be done before Last Call - don't want to do before final changes
08:11:23 [pfps]
Ivan: What about Theorem 1
08:11:41 [pfps]
IanH: Up for discussion later
08:11:50 [pfps]
IanH: Some issues related to RL
08:12:00 [IanH]
08:12:04 [pfps]
Subtopic: Primer
08:12:23 [pfps]
Bijan: I'm waiting for the other documents to stabilize
08:12:41 [pfps]
Bijan: I might want to change the example - traditional families might be controversial
08:13:17 [pfps]
Ivan: For me the example works - I propose not to change unless there are major objections
08:13:52 [IanH]
08:14:27 [pfps]
Sandro: Stay biological - social is controversial
08:15:09 [pfps]
Ivan: Turtle examples are not nice - I will work on them
08:15:50 [pfps]
Bijan: I can't commit to Primer before end of year
08:16:33 [pfps]
Ivan: What is the status of the primer - rec track vs note - undetermined so far
08:16:49 [pfps]
Ivan: What about profiles in primer?
08:16:53 [IanH]
08:17:13 [pfps]
Bijan: As little as possible - bulks up the primer too much
08:17:22 [IanH]
08:17:35 [pfps]
Ivan: How about using the same example for all profiles?
08:17:43 [pfps]
Bijan: Could be a good idea
08:17:54 [pfps]
Ivan: Appendices?
08:18:12 [pfps]
Bijan: Profiles in text - non-starter - overwhelming
08:18:24 [IanH]
08:18:26 [pfps]
Bijan: Profiles in appendices - better
08:18:30 [pfps]
Ivan: More useful
08:18:41 [pfps]
Bijan: Let's try one of them
08:18:45 [pfps]
Ivan: I'll try RL
08:19:35 [Zhe]
Ivan I can help you if you need anything
08:19:56 [IanH]
08:20:40 [pfps]
Christine: Primer is similar to Ontology Development 101, which was useful
08:23:14 [IanH]
08:23:22 [pfps]
Christine: I don't like the Manchester Syntax - it is frame-like and uses "fact" - may lead to misunderstanding
08:24:29 [pfps]
Bijan: The Primer just uses the majorly-used syntaxes - We used Manchester syntax initially so it comes first
08:24:32 [IanH]
08:25:33 [IanH]
08:26:25 [pfps]
IanH: We will discuss status and schedule later in the F2F.
08:27:26 [pfps]
Bijan: There is some perspective-specific stuff in the primer (that can be removed from the presentation)
08:27:32 [Elisa]
Latest version of the QRG (wiki) is at
08:27:37 [pfps]
Subtopic: Quick Reference Guide:
08:28:16 [pfps]
pfps: Agenda has pointer to most recent version
08:28:21 [IanH]
08:28:26 [pfps]
Ivan: QRG has changed tremendously
08:28:42 [pfps]
Elisa: Yes it did change a lot, and it changed again just recently
08:28:50 [pfps]
Ivan: QRG looks good
08:29:08 [pfps]
Elisa: We took a recommendation from pfps to reorganize
08:29:18 [pfps]
Elisa: Not everything is hyperlinked
08:29:29 [IanH]
08:29:58 [ivan]
08:30:01 [pfps]
Elisa: Intent is to hyperlink everything (functional syntax, RDF syntax, etc.)
08:30:17 [pfps]
Elisa: Might also link to Primer
08:30:29 [pfps]
Elisa: Might require anchors in other documents
08:30:43 [pfps]
Elisa: Still want a two-page print version from this structure
08:30:59 [pfps]
Elisa: Also want a page for the profiles
08:31:18 [bijan]
I like this a lot!
08:31:36 [pfps]
Elisa: Examples - we might not keep them but instead link to Primer
08:31:39 [IanH]
08:31:39 [bijan]
Or link to the syntax, which has examples for every feature
08:32:32 [IanH]
08:32:56 [pfps]
Elisa: We want feedback on structure, later sections need more review
08:33:17 [FabGandon]
FabGandon has joined #owl
08:33:22 [IanH]
08:33:22 [bijan]
Note: Old documents had a similar multi-docuoment index:
08:33:26 [bijan]
But this is much nicer
08:33:35 [pfps]
Ivan: I like it
08:33:52 [bijan]
And is much better than:
08:33:56 [pfps]
Ivan: What should the third column link to?
08:34:17 [pfps]
Elisa: We are not sure - I think semantics
08:34:30 [IanH]
08:34:44 [pfps]
Ivan: Mapping document is just a table - so not good to link to it - semantics is better
08:35:44 [pfps]
Michael: One problem is that RDF semantics doesn't have the "syntax"
08:35:54 [pfps]
IanH: Semantic isn't great to link to
08:36:18 [pfps]
Elisa: Might link to Primer instead - we may try some things
08:36:50 [IanH]
08:36:53 [pfps]
Bijan: Neither RDF mapping nor RDF semantics is useful to link to
08:36:57 [IanH]
ack ivan
08:37:08 [pfps]
Ivan: perhaps linking to primer is best
08:37:30 [pfps]
Bijan: Primer is not comprehensive but could serve, perhaps with minor changes
08:38:53 [IanH]
08:39:00 [pfps]
Christine: QRG is most useful as initial point of contact
08:39:31 [pfps]
Christine: QRG is too terse
08:40:17 [pfps]
Christine: LInk to requirements document instead?
08:40:56 [IanH]
08:41:09 [pfps]
Evan: What about linking from Recommendations to Notes
08:41:21 [pfps]
Ivan: Not a good idea
08:41:28 [IanH]
08:41:29 [pfps]
Bijan: I don't see a problem - just need to be careful
08:41:48 [pfps]
Ivan: Need to refer to stable documents
08:41:55 [pfps]
Bijan: I like the document
08:42:18 [pfps]
IanH: QRG is getting close to being done, still needs work
08:42:44 [pfps]
Bijan: Publish as working draft at last call, even if not done
08:43:10 [Zakim]
08:43:16 [pfps]
Subtopic: Requirements
08:43:49 [IanH]
08:44:17 [pfps]
Christine: I think that requirements is close to done - I would make changes - may need changes based on F2F discussion
08:44:37 [pfps]
Christine: There have been several reviews - Bijan, Jie, Elisa
08:44:48 [pfps]
Christine: Only Bijan had major comments
08:44:57 [IanH]
08:45:45 [pfps]
Christine: Addressing Bijan's comments needs input from WG
08:46:21 [pfps]
Christine: There are some conflicting reviews
08:46:48 [pfps]
Christine: Almost all done - changes needed in response to outstanding comments
08:47:15 [pfps]
Christine: Major decision is whether to cut chunks out
08:47:25 [pfps]
Ivan: I like Section 5
08:47:59 [pfps]
Ivan: What does the button do?
08:48:23 [pfps]
Ivan: Oh, I see -
08:48:47 [pfps]
Evan: Need feedback on what do to with the document
08:49:02 [pfps]
Evan: One possibility is to split into two
08:49:04 [Zhe]
+1 to Evan
08:50:20 [pfps]
Rinke: Large fraction of HCLS use cases - how about recategorizing them?
08:51:26 [pfps]
IanH: Need to discuss this document later
08:51:50 [pfps]
Christine: Suggest to move features to Quick Reference Guide
08:53:45 [sandro]
09:17:18 [ivan]
alexandre passant
09:17:58 [sandro]
holger stezhorm
09:18:00 [pfps]
holger stenzhorl
09:18:09 [pfps]
scot marshall
09:18:22 [ivan]
09:18:30 [sandro]
09:18:46 [Rinke]
Rinke has joined #owl
09:18:47 [bmotik]
Blaz Novak
09:18:58 [pfps]
Subtopic: Manchester Syntax
09:19:35 [pfps]
pfps: up to date - perhaps one or two issues that might affect it
09:20:17 [pfps]
pfps: there have 2.5+? reviews - one substantive comment
09:21:37 [pfps]
christine: what about mapping from functional to manchester?
09:21:56 [pfps]
pfps: responded with comment that the mapping is "trivial" - comment remains in document
09:22:07 [pfps]
Subtopic: Internationalized String Spec
09:22:34 [pfps]
Boris: Still waiting on Axel Polares for built-in functions (wanted by RIF)
09:22:41 [alanr]
alanr has joined #owl
09:23:58 [pfps]
jeremy: should refer to RDF 4647 as well as 4646 - which may result in changes
09:24:11 [pfps]
09:24:33 [pfps]
boris: what is the impact
09:24:51 [pfps]
jeremy: may need to change matching
09:25:06 [pfps]
IanH: plan / schedule reviewing?
09:25:22 [pfps]
Ivan: needs to be at least a WD by last call
09:25:32 [pfps]
Ivan: what is RIF status?
09:25:46 [pfps]
Sandro: waiting for Axel's changes
09:26:08 [pfps]
IanH: We need to wait for changes
09:26:19 [Achille]
Achille has joined #owl
09:26:32 [sandro]
09:26:39 [pfps]
jeremy: also review I18N
09:26:40 [sandro]
I18N == "Internationalization"
09:27:10 [pfps]
Bijan: we should push a FPWD ASAP - it blocks us
09:27:26 [pfps]
Bijan: What does CR mean for this?
09:28:03 [pfps]
pfps: if we don't care about built-ins why not push for our approval
09:28:37 [pfps]
Ivan: The CR criteria are the purview of the OWL WG and the RIF WG
09:29:05 [pfps]
Ivan: There could be different CR exit criteria from the rest of our spec
09:29:59 [pfps]
Boris: I sent Axel a message
09:30:15 [pfps]
Bijan: let's push the document even without the built ins
09:30:20 [pfps]
IanH: We don't need them at all
09:30:24 [pfps]
Ivan: RIF wants them
09:30:57 [pfps]
Boris: RIF thought that the old version was lopsided (as it had facets but not built-ins)
09:31:27 [pfps]
Boris: they may not sign off without built-ins
09:31:58 [pfps]
Alan: a WG can have open areas - a section with a missing bit is OK
09:32:36 [pfps]
Ivan: we vote to publish ASAP even if there is a missing section
09:33:13 [pfps]
Alan: include editor's note in FPWD
09:33:32 [pfps]
IanH: can we push now?
09:34:09 [pfps]
IanH: tomorrow we can vote on this
09:34:56 [pfps]
Bijan: this has come from us, and we discussed it, so we don't really need *more* review
09:35:46 [pfps]
Bijan: what about patent review? this means that we *need* FPWD at least 90 days before end
09:36:09 [pfps]
ScribeNick: bernardo
09:36:41 [bcueencagrau]
IanH: outstanding issues
09:36:43 [Blaz]
Blaz has joined #owl
09:37:01 [bcueencagrau]
IanH: first issue, punning (Isuue 114)
09:37:29 [bcueencagrau]
bmotik: the issues with annotations are orthogonal to punning
09:37:48 [bcueencagrau]
bmotik: we will add a section on punning, where we will explain what punning is by example
09:38:28 [bcueencagrau]
bcuencagrau: we will explain punning as different ``views'' over the same URI
09:38:40 [bcueencagrau]
bmotik: there was a proposal concerning annotations
09:39:45 [bcueencagrau]
bmotik: annotations do not have semantics in DLs
09:40:02 [bcueencagrau]
bmotik: instead of attaching annotations to entities we would attach annotations to URIs
09:40:13 [wallace]
wallace has joined #owl
09:40:17 [bcueencagrau]
bmotik: this would make the mapping to RDF easier
09:40:37 [ivan]
09:40:44 [bcueencagrau]
alanr: talking of annotations as being a URI might be confusing
09:41:30 [bcueencagrau]
alanr: there is a connection between annotations and punning
09:41:54 [bcueencagrau]
ianH: could we think of this issue as a bug fix?
09:42:20 [bcueencagrau]
alanr: there may also be a problem with anonymous individuals and literals
09:42:49 [IanH]
09:43:08 [IanH]
ack ivan
09:43:17 [bcueencagrau]
bmotik: the only change is that the values of annotations will change from entities to URIs
09:43:56 [alanr]
09:44:08 [bcueencagrau]
ivan: this idea of `view' over entities is irrelevant from the OWL Full point of view
09:44:57 [bcueencagrau]
\nick bernardo
09:45:05 [alanr]
1) URI not the greatest name
09:45:34 [alanr]
2) Explain missing individual sameAs => extensions equivalent
09:46:01 [bernardo]
bparsia: there will be a section on punning/metamodeling
09:46:36 [IanH]
09:46:40 [IanH]
ack alanr
09:47:47 [bernardo]
alanr: we should make in the document a clear distinction between the OWL Full and DL views
09:48:05 [bernardo]
ianH: where are we concerning this issue?
09:49:08 [bernardo]
ianH: we are essentially fixing a bug
09:49:27 [bernardo]
bmotik: there is an email with a proposal that already acknowledges this bug
09:49:51 [bernardo]
bmotik: resolution would involve adding the new section and make the sall change on annotations
09:50:15 [bernardo]
bparsia: let us resolve this issue as in Borise's proposal
09:50:47 [bmotik]
09:50:48 [bernardo]
IanH: modulo editorial issues we should have a clear proposal to resolve
09:51:49 [bmotik]
09:51:55 [cgolbrei]
cgolbrei has joined #owl
09:52:49 [IanH]
PROPOSED: resolve issue 114 as per and subsequent thread
09:53:21 [bmotik]
+1 (Oxford)
09:53:31 [bernardo]
+1 (Oxford)
09:53:59 [bijan]
+1 (Manchester)
09:54:03 [Rinke]
+1 (UvA)
09:54:06 [bernardo]
sandro: is the resolution proposal in the email itself or in the thread?
09:54:25 [sandro]
Ian: There is still some editorial work to be done, which might open up related issues.
09:54:26 [bernardo]
ianH: we have agreed that there will be some editorial issues involved
09:54:28 [MarkusK_]
+1 (FZI)
09:54:30 [pfps]
+1 (ALU)
09:54:34 [alanr]
+1 (Science Commons)
09:54:34 [sandro]
+1 (W3C)
09:54:38 [Zhe]
09:54:45 [IanH]
+1 (Oxford)
09:54:46 [Achille]
+1 (IBM)
09:54:53 [cgolbrei]
+1 (uvsq)
09:55:01 [wallace]
0 (NIST)
09:55:01 [sandro]
Zakim, who is here?
09:55:01 [Zakim]
On the phone I see Riviera_B, Zhe (muted)
09:55:02 [Zakim]
On IRC I see cgolbrei, wallace, Blaz, Achille, alanr, Rinke, FabGandon, schneid, bijan, bmotik, Zhe, MarkusK_, sandro, bernardo, IanH, RRSAgent, Zakim, pfps, ivan, trackbot
09:55:33 [IanH]
RESOLVED: resolve issue 114 as per and subsequent thread
09:55:46 [sandro]
RRSAgent, pointer?
09:55:46 [RRSAgent]
09:56:13 [bernardo]
Issue 134
09:56:29 [bernardo]
alanr: we have asked Elisa for feedback
09:56:52 [bernardo]
Evan: speaking for Elisa. The person who was our contact is moving at the moment
09:57:18 [bernardo]
alanr: will evan have time to look into it?
09:57:39 [bernardo]
evan: the problem is getting to work the tools needed
09:57:45 [IanH]
09:57:54 [bernardo]
evan: I can talk with Peter Haase
09:58:00 [IanH]
09:58:07 [bernardo]
evan: Peter can provide the tool
09:58:33 [bernardo]
alanr: first issue: what is the impact of the metamodel in the current docs?
09:58:52 [bernardo]
ianH: we need to figure out whether we want to have it at all
09:59:15 [bernardo]
bmotik: concerning accessibility, it is from someversion ofEclipse
09:59:36 [bernardo]
bmotik: the diagrams are in an IBM format
09:59:47 [IanH]
09:59:51 [bernardo]
bmotik: the metamodel can be used to a large extent using Eclipse
10:00:34 [bernardo]
bmotik: the metamodel is a representation of the diagrams in a machine readable, formal way
10:00:44 [IanH]
10:01:03 [bernardo]
bmotik: layout-related information is not part of the metamodel, but on top of it
10:01:15 [bernardo]
bmotik: this extra stuff is in the IBM format
10:01:29 [bernardo]
bmotik: the structure can be viewed using Eclipse
10:01:50 [bernardo]
bmotik: Eclipse 3.4. with EMF on
10:01:59 [IanH]
10:02:07 [bernardo]
bmotik: it does not have a diagram capability, though
10:03:05 [bernardo]
bparsia: if the metamodel does involve the layout information, but this is largely unimportant
10:03:18 [bernardo]
IanH: do we need to have this as a part of the spec?
10:03:38 [bernardo]
alanr: I would like to have a clear metamodel in a machine-readable format
10:04:14 [bernardo]
alanr: the content of the document Met describing the metamodel has changed
10:04:22 [bernardo]
bmotik: I disagree
10:04:25 [IanH]
10:04:56 [bernardo]
evan: I like to look at the diagrams
10:05:24 [alanr]
10:05:52 [bernardo]
evan: the situation is not that bad. There's other tools that can be used for layout
10:06:29 [IanH]
10:06:30 [bernardo]
evan: we have public domain tools that allow us to look at the metamodel. A separate question is whether it conforms to MOF
10:07:21 [bernardo]
bparsia: the way you would do it is to read the text, look at the diagrams and even click on them and get some sort of code (e.g. javascript) that represents the diagram
10:08:27 [bernardo]
bmotik: metamodel is about connectivity
10:08:35 [IanH]
10:08:50 [bernardo]
bmotik: in IBM RSA you can draw the diagrams and get the metamodel on the fly
10:09:12 [IanH]
10:09:16 [IanH]
ack alanr
10:09:19 [bernardo]
bmotik: we have to insist on total MOF compliance
10:09:41 [bernardo]
alanr: what is the benefit of having this metamodel?
10:10:10 [bernardo]
bmotik: we want to have a very precise statement saying what is the structure of OWL
10:10:22 [bernardo]
bmotik: this would be described by the metamodel
10:10:41 [bernardo]
bmotik: precision is provided by the metamodel, and that alone is enough to justify it
10:10:43 [IanH]
10:10:46 [IanH]
10:11:07 [bernardo]
bmtik: from a practical point of view, people could generate classes directly from the metamodel
10:11:24 [bernardo]
bmotik: also how to transform between metamodels
10:11:37 [alanr]
q+ to ask a few follow questions
10:11:50 [bernardo]
bmotik: but the important thing is to have a precise specification
10:12:08 [bernardo]
bparsia: I am ok with the text, the metamodel would be good but not a must
10:12:26 [alanr]
q+ to propose it not be critical path to lc
10:12:34 [bernardo]
bparsia: the benefit would be in people using it
10:13:16 [bernardo]
jeremy: is the extension of the document with the machine-readable metamodel intended to be normative or informative?
10:13:25 [bernardo]
Ianh: that is a question
10:14:09 [bernardo]
bparsia: what we mean is that, if we have a conflict between the text and the metamodel, the metamodel wins
10:14:27 [bernardo]
ivan: this is a lot of work, and probably not a priority
10:14:32 [IanH]
10:14:46 [bernardo]
10:15:21 [bernardo]
christine: I am not clear what boris means with ``precision''
10:15:56 [bernardo]
christine: not clear what is the real benefit
10:16:15 [bijan]
+1 I have good confidence in the current document
10:16:18 [bernardo]
alanr: if it is normative, then we will have a lot of work
10:16:22 [bijan]
But I also think the metamodel is useful
10:16:40 [bernardo]
alanr: I suggest it to become a note
10:17:48 [bernardo]
ianH: nobody thinks that there is no way we shouldn't have it
10:18:45 [bernardo]
bmotik: sometimes English introduces ambiguity. Ambiguity is bad for implementors
10:19:10 [cgolbrei]
but we do have the diagrams in the syntax
10:22:10 [sandro]
Sandro: As long as it doesn't affect implementations, it's okay after Last Call
10:22:39 [sandro]
Jeremy: There is a risk to adding a new master version of the spec, after last call.
10:22:50 [sandro]
Bijan: Yes, it's a risk, but it's doable if necessary.
10:23:20 [bernardo]
bmotik: one question, how bout changing ``UML'' into MOF?
10:24:41 [sandro]
Bernardo: If we have this formal metamodel, then what would the review consist of? Someone checking english text vs formal model? Thousands of people look at this, then send their reports to the list?
10:24:56 [bmotik]
10:25:11 [bernardo]
Achille: the diagrams we have are supposed to play the role of the MOF metamodel
10:25:28 [bernardo]
achille: the spac has already been carefully reviewed
10:25:54 [bernardo]
achille: we could probably add a simple note stating the priority of the diagrams over the text
10:26:20 [bernardo]
achile: If we do not have a formal description, I wouldn't care that much
10:26:42 [bernardo]
IanH: so, there seems to be consensus that this is not in our critical path
10:26:54 [IanH]
10:27:05 [bernardo]
10:27:26 [bernardo]
MarkusK: if there is work to be done Peter Haase can contribute
10:27:53 [IanH]
ack alanr
10:27:53 [Zakim]
alanr, you wanted to ask a few follow questions and to propose it not be critical path to lc
10:28:00 [IanH]
ack bernardo
10:28:15 [bernardo]
bijan: we could still have it, and not lose this work
10:28:40 [bernardo]
bparsia: whether we have it as an appendix or somewhere else doesn't really matter
10:28:40 [IanH]
10:29:17 [bernardo]
alanr: we could then, as Boris suggests, simply make an editorial change
10:29:50 [cgolbrei]
link to MOF please ?
10:29:54 [MarkusK_]
pointer to MOF on the wiki:
10:29:57 [sandro]
Jeremy: If they MOF is normative, then what normative documents are are needed to understand it?
10:30:02 [bernardo]
bparsia; currently we do not have any statement saying that the diagrams have priority over the text
10:30:12 [bernardo]
10:31:05 [bernardo]
sandro: it is a good practice to have several normative ways, because it can help identifying bugs
10:31:53 [sandro]
Jeremy: While the OWL1 test cases are normative, it's stated that the Semantics rules.
10:32:10 [bernardo]
bmotik: how about changing the first sentence of Section 2.1
10:32:55 [bernardo]
bmotik: I want to say that the diagrams are intended to define the structure of the language
10:33:22 [bernardo]
alanr: I do not understand the difference between normative and controlling
10:34:00 [bernardo]
alanr: the question is whethr among two normative representations, there is one that is controlling
10:34:30 [bernardo]
bparsia: we need a controlling view for interoperability issues
10:34:42 [sandro]
+1 Bijan: it's best to pick a controlling normative version, so you get more interoperability.
10:34:53 [sandro]
Bijan: You do less harm if you have people converge.
10:35:20 [sandro]
Peter: if the controlling one is flat-out-broken, then people will revolt and use the other one.
10:35:23 [bernardo]
peter: agree with bijan
10:35:53 [bernardo]
ianH: the conclusion is that we say what is the controlling representation
10:36:07 [sandro]
Ian: I'm hearing that people do want to pick a controlling one.
10:36:29 [bernardo]
ivan: we have two candidates
10:36:32 [IanH]
10:36:47 [bernardo]
ivan: functional + text versus diagrams
10:36:47 [sandro]
10:36:50 [sandro]
10:37:14 [bernardo]
bparsia: 1) diagrams plus english
10:37:22 [bernardo]
2) grammar + english
10:37:28 [IanH]
10:37:34 [bernardo]
3) metamodel + english
10:37:56 [bernardo]
bmotik: I agree these are the candidates
10:38:24 [bernardo]
achille: there is an ambiguity when we talk about ``text''
10:38:41 [bernardo]
bparsia: this is a presentation issue
10:39:04 [bernardo]
mschneider: diagrams plus English tells us the whole story?
10:39:28 [bernardo]
bmotik: you could use either 1) or 2)
10:40:25 [bernardo]
christine: what is the difference between 1) and 3)?
10:40:45 [bernardo]
christine: what about diagrams plus grammar?
10:40:52 [bernardo]
bmotik: this is not sufficient
10:40:52 [alanr]
alanr has joined #owl
10:41:05 [bernardo]
IanH: we should have a straw poll
10:42:35 [bernardo]
IanH: it seems that the vast majority liked 1) and 3)
10:44:29 [bernardo]
bparsia: we should have a sentence at the beginning of the syntax doc explaining this
10:45:06 [bernardo]
alanr: we should add information saying how to read the diagrams
10:45:50 [bernardo]
bparsia: we need to have a pointer to the formal meaning of the diagrams
10:46:36 [bernardo]
jeremy: there is an issue concerning people who cannot see the diagrams
10:47:09 [bernardo]
bparsia: we should have an official resolution about the controlling text
10:47:36 [bernardo]
proposal: the diagrams plus the supportive text are the definitive specification
10:48:32 [IanH]
PROPOSAL: the diagrams plus the supportive text are the definitive specification and there will be a normative reference for the diagrams
10:49:01 [IanH]
ROPOSAL: the diagrams plus the supportive text are the definitive specification for the language and there will be a normative reference for the diagrams
10:49:24 [IanH]
PROPOSAL: the diagrams plus the supportive text are the definitive specification for the language and there will be a normative reference for the diagrams
10:49:26 [pfps]
+1 (ALU)
10:49:29 [Achille]
+1 (IBM)
10:49:32 [bernardo]
+1 (Oxford)
10:49:32 [wallace]
+1 (NIST)
10:49:32 [bijan]
+1 (Manchester)
10:49:35 [IanH]
+1 (Oxford)
10:49:37 [Zhe]
10:49:37 [bmotik]
+1 (Oxford)
10:49:42 [MarkusK_]
+1 (FZI)
10:49:47 [alanr]
0 (Science Commons)
10:49:49 [Rinke]
+1 (UvA)
10:50:00 [IanH]
RESOLVED: the diagrams plus the supportive text are the definitive specification for the language and there will be a normative reference for the diagrams
10:50:01 [sandro]
10:50:18 [ivan]
10:50:30 [cgolbrei]
+1 (uvsq)
10:50:50 [bernardo]
bparsia: I already have an action for checking about accessibility
10:51:52 [bernardo]
bparsia: finding alternative text for the diagrams that is accessibility-friendly is very hard
10:52:50 [bernardo]
bparsia: we should have some text, because it is more useful for disabled people
10:53:58 [bernardo]
IanH: should we move on?
10:54:19 [bernardo]
IanH: we can skip it and it is not in our critical path for last call
10:57:19 [bernardo]
bmotik: we could close the issue
10:57:21 [sandro]
Ian: could close and say it'll be a note, if anything.
10:58:17 [bernardo]
IanH: that would mean that the MOF XML will not be included in Syntax
10:59:39 [bernardo]
evan: I wouldn't like o close it like that
11:00:10 [bernardo]
bparsia: if we had a great MOF, we could always revise
11:00:38 [sandro]
PROPOSED: Close Issue-134, saying we don't expect to have a MOF metamodel in a Rec-Track document. Maybe a Note someday.
11:00:47 [bmotik]
+1 (Oxford)
11:00:51 [ivan]
11:00:51 [Rinke]
+1 (UvA)
11:00:51 [pfps]
+1 (ALU)
11:00:53 [bernardo]
+1 (Oxford)
11:00:54 [wallace]
0 (NIST)
11:00:54 [Achille]
0 (IBM)
11:00:55 [MarkusK_]
+0 (FZI)
11:00:56 [IanH]
+1 (Oxford)
11:00:57 [Zhe]
11:01:03 [sandro]
+1 (W3C)
11:01:11 [bijan]
+1 (Manchester)
11:01:14 [alanr]
11:01:37 [cgolbrei]
11:01:50 [sandro]
RESOLVED: Close Issue-134, saying we don't expect to have a MOF metamodel in a Rec-Track document. Maybe a Note someday.
11:09:48 [Zakim]
11:40:09 [bmotik]
bmotik has joined #owl
11:51:35 [Rinke]
Rinke has joined #owl
11:59:00 [IanH]
IanH has joined #owl
12:07:33 [IanH_]
IanH_ has joined #owl
12:07:42 [schneid]
schneid has joined #owl
12:09:32 [msmith]
msmith has joined #owl
12:09:47 [sandro]
scribe: schneid
12:09:49 [schneid]
Topic: 142
12:09:54 [schneid]
Topic: Issue-142
12:10:28 [schneid]
bmotik: there are actually three issues
12:11:00 [schneid]
alanr: about the sketch proof of thereme 1
12:11:43 [alanr]
alanr has joined #owl
12:12:37 [schneid]
ianh: looked at proof and looked reasonable
12:13:00 [schneid]
alanr: proposal is to close issue
12:13:17 [schneid]
bparsia: i have reviewed it too, and was fine
12:13:58 [alanr]
Proposed: Resolve Issue 142 by Addition of a proof sketch to the profiles document
12:14:05 [pfps]
+1 (ALU)
12:14:17 [bmotik]
12:14:19 [Achille]
Achille has joined #owl
12:14:24 [MarkusK_]
12:14:25 [wallace]
wallace has joined #owl
12:14:27 [bernardo]
12:14:27 [Rinke]
12:14:29 [Achille]
12:14:30 [IanH_]
12:14:30 [alanr]
12:14:37 [wallace]
12:14:38 [pfps]
Bijan: +1
12:14:52 [sandro]
12:15:07 [alanr]
RESOLVED: Resolve Issue 142 by Addition of a proof sketch to the profiles document
12:15:23 [schneid]
Topic: Issue-145
12:16:09 [schneid]
sandro: any format we publish on the web must have an own mime type
12:17:10 [schneid]
we have to tell the ITF
12:17:19 [schneid]
sandro: we have to tell the ITF
12:17:51 [schneid]
sandro: my example that I sent by mail was what I sent for RIF
12:18:10 [ivan]
ivan has joined #owl
12:18:46 [schneid]
alanr: so this is about OWL/XML syntax, functional syntax, and manchester syntax
12:19:18 [schneid]
ivan: we definitely need it for functional and manchester
12:19:24 [schneid]
sandro: we can use text/plain
12:21:22 [schneid]
alanr: another consideration is predictability
12:22:06 [schneid]
bijan: one cannot rely on mimetypes when one wants to disambiguate
12:22:27 [sandro]
W3C Process document on this:
12:22:59 [schneid]
alanr: two extreme options: lowest = application/xml and text/plain, maximum is a new mime type for each
12:23:08 [sandro]
My e-mail on this for RIF, giving an example of how much to do:
12:24:05 [bmotik]
bmotik has joined #owl
12:24:32 [schneid]
bparsia: does turtle or n3 has a mime type
12:25:25 [schneid]
sandro: we should have mime type iff particular syntax is intended as a definitive exchange syntax on the web (his opinion)
12:27:09 [schneid]
ivan: reformulate what sandro sais: if syntax rec track, then pressure on us to give mime type higher
12:27:26 [schneid]
alanr: let's straw poll
12:28:01 [schneid]
sandro: generic mime types do sort of say: "don't put this stuff on the web"
12:28:43 [schneid]
alanr: if a syntax isn't first class, then we shouldn't do anything about it.
12:29:44 [schneid]
bparsia: looks like that wg thinks that we don't have the resources to bring manchester syntax on the rec track
12:30:18 [schneid]
ivan: we can still decide to give mime type to manchester syntax
12:30:37 [Zakim]
12:31:54 [sandro]
Sandro: I think we need to require all OWL implementations to parse all "first-class" serializations.
12:32:06 [sandro]
Sandro: Otherwise we fragment the OWL world.
12:32:56 [schneid]
ivan: we have mime types for RDF/XML and turtle
12:33:13 [schneid]
bparsia: wasn't hard to come up with converters
12:33:38 [schneid]
bmotik: is it realistic to require tools to support all the different syntaxes?
12:34:13 [schneid]
achille: not clear why is it our responsibility to give mime type to manchester syntax
12:34:55 [schneid]
bparsia: pfps and the wg did much of the work on the manchester draft
12:36:56 [schneid]
ivan: from rdf point of view, system only has to understand rdf/xml; turtle is around, but not needed to be supported officially
12:37:13 [sandro]
Sandro: We have a duty to tell people which syntaxes they can publish in and assume that conformant OWL consumers will understand.
12:37:26 [schneid]
ivan: because manchester is not rec, we cannot say that tools have to support it
12:38:07 [schneid]
ivan: mime type discussion should be primarily be on owl/xml, because this is on rec track
12:38:42 [schneid]
ivan: for functional, it has to be discussed, whether it is meant to be put into a file and send to the web, then mime too
12:39:24 [wallace]
can we call the question on this?
12:39:41 [schneid]
pfps: case of owl 1, it looks that the only normative exchange syntax was rdf/xml, but could also have been the abstract syntax
12:41:11 [schneid]
bparsia: in owl 2 we should not have one normative exchange syntax [fixme]
12:41:40 [schneid]
sandro: in owl 2, the normative exchange syntax should be rdf/xml, as in owl 1
12:43:12 [schneid]
sandro: if we have more than one mandatory exchange syntax, then we have a problem
12:43:51 [schneid]
pfps: i would at least see the functional syntax as an official syntax, and would also prefer manchester to be official
12:44:22 [sandro]
sandro: ... where owl/xml is allowed, using GRDDL.
12:44:55 [sandro]
conformant provider --- provides in a normative stynax
12:46:10 [schneid]
bparsia: if we would use n-triple syntax only for test cases but would say that n-triple is not an official syntax, than this would be a problem
12:47:12 [schneid]
sandro: if someone publishes n-triple, than he can do so, but has also to give the rdf/xml version
12:48:27 [schneid]
pfps: more and more people produce rdf document not in official syntax, which is a good thing
12:49:47 [schneid]
alanr: peter, is what you say about content negotiation?
12:51:14 [schneid]
alanr: let's settle on a few things: mime does not need normativity of syntax...
12:51:32 [sandro]
Alan: Let's do MIME types for all the syntaxes, to enable content negotiation.
12:54:07 [schneid]
alanr: strawpoll on mimetype or not mimetype
12:54:53 [alanr]
straw poll: Mime types for functional, xml, manchester syntaxes?
12:54:57 [ivan]
12:55:00 [MarkusK_]
12:55:05 [Achille]
12:55:09 [Rinke]
12:55:09 [sandro]
1 -- because con-neg is useful
12:55:10 [alanr]
pfps +1
12:55:13 [bmotik]
12:55:16 [Zhe]
12:55:17 [bernardo]
12:55:20 [schneid]
+0 (it's perhaps a lot of work)
12:55:21 [cgolbrei]
12:55:24 [alanr]
bijan: -1
12:55:36 [wallace]
12:55:39 [alanr]
12:56:53 [sandro]
alan: strong majority in favor of more mime types.
12:57:33 [Hyunjeong]
Hyunjeong has joined #owl
12:57:50 [schneid]
m_schnei: I fear that this will be a lot of work on WG
12:57:54 [schneid]
ivan: no, it's trivial
12:58:39 [schneid]
alanr: will someone of the against-voters be opt against in real vote?
12:59:18 [schneid]
achille: i am not comfortable, and perhaps would vote against
13:00:37 [schneid]
markusk: giving a functional syntax doesn't change the state of the respective syntax, but likes to see that people can distinguish between syntaxes
13:01:49 [schneid]
alanr: explains difference between having mime type and normativity
13:02:14 [schneid]
ianh: last telco, same discussion in mime types, and normativity was not a topic
13:02:22 [IanH_]
13:02:26 [Achille]
achille: I am opposing the use of functional and manchester syntaxes as standard exchange formats.
13:02:53 [schneid]
bparsia: i have to first talk to uli before i can decide wheter opt against or not
13:03:50 [schneid]
achille: main concern is to clarifiy that from having mime type no assertion about normativity of exchange syntax follows
13:04:33 [pfps]
pfps has joined #owl
13:04:46 [schneid]
ivan: it seems that unless uli formally objects, we can proceed on mime type topic
13:05:31 [schneid]
bparsia: uli is in favour of progress
13:06:13 [schneid]
alanr: what mime types?
13:06:28 [schneid]
ivan: obvious for OWL/XML: owl+xml
13:06:56 [schneid]
sandro: does OWL 1 XML syntax had mime type
13:07:02 [schneid]
pfps: no, no mime type
13:07:06 [alanr]
PROPOSAL: We will defined mime types for the functional syntax, manchester syntax, and owl syntax. The mime type for the owl syntax will be application/owl+xml
13:07:32 [alanr]
PROPOSAL: We will define mime types for the functional syntax, manchester syntax, and owl syntax. The mime type for the owl syntax will be application/owl+xml
13:07:50 [pfps]
+1 (ALU)
13:07:51 [wallace]
13:07:58 [IanH_]
+1 (Oxford)
13:07:59 [Achille]
0 (IBM)
13:08:34 [cgolbrei]
0 (uvsq)
13:08:46 [alanr]
PROPOSAL: We will define mime types for the functional syntax, manchester syntax, and owl syntax. The mime type for the owl syntax will be application/owl+xml. This does not speak to any of these being normative exchange syntaxes
13:08:48 [Zhe]
13:08:51 [Rinke]
+1 (UvA)
13:08:54 [pfps]
+1 (ALU)
13:08:55 [alanr]
+1 (Science Commons)
13:08:55 [wallace]
0 (NIST)
13:08:59 [IanH_]
+1 (Oxford)
13:09:03 [ivan]
1 (W3C)
13:09:05 [Achille]
0 (IBM)
13:09:11 [schneid]
+1 (FZI)
13:09:54 [cgolbrei]
+1 (uvsq)
13:10:16 [schneid]
alanr: observes that no vote from manchester
13:10:25 [alanr]
RESOLVED: We will define mime types for the functional syntax, manchester syntax, and owl syntax. The mime type for the owl syntax will be application/owl+xml. This does not speak to any of these being normative exchange syntaxes
13:10:27 [schneid]
bparsia: we have to wait for uli to object or not
13:10:40 [MarkusK_]
MarkusK_ has joined #owl
13:12:02 [schneid]
alanr: we need people who work on the mime types
13:12:59 [alanr]
action: Ivan to propose mime types for functional and manchester syntaxes
13:12:59 [trackbot]
Created ACTION-233 - Propose mime types for functional and manchester syntaxes [on Ivan Herman - due 2008-10-30].
13:16:18 [MarkusK_]
MarkusK_ has left #owl
13:16:23 [MarkusK_]
MarkusK_ has joined #owl
13:17:00 [schneid]
Topic: Datatypes
13:17:07 [schneid]
SubTopic: n-ary Datatypes
13:17:22 [IanH_]
Can we have a volunteer to scribe in the next session please.
13:18:18 [msmith]
msmith has joined #owl
13:18:23 [bernardo]
bernardo has joined #owl
13:18:25 [schneid]
bparsia: uli want's 1/3 be representable
13:18:50 [schneid]
alanr: what is the connection between n-aries and rationals?
13:20:17 [schneid]
bparsia: xsd:decimals don't represent all possible rationals, and we want solvability, and want to write down the literals
13:21:17 [schneid]
bmotik: satisfiability problem is actually solved by owl:real
13:21:45 [ivan_]
ivan_ has joined #owl
13:22:07 [schneid]
alanr: is "rational" issue dependent on "n-ary" issue?
13:23:03 [schneid]
alanr: we don't have n-aries in main spec, but could have it in a side spec. do we have this additional spec ?
13:23:18 [schneid]
alanr: where do we stand with n-ary spec?
13:23:55 [schneid]
bparsia: we have support in syntax, in semantics, stratification (3 levels)
13:24:16 [schneid]
alanr: (to bparsia) to you have a document?
13:24:31 [schneid]
bparsia: we have syntax and semantics, so essentially yes
13:25:02 [schneid]
alanr: we need reviewers
13:27:24 [schneid]
bparsia: 3 levels: plain comparison, scaled comparison, linear comparison
13:28:05 [schneid]
bparsia: gives examples...
13:28:21 [schneid]
... plain comparsion, "h > w"
13:28:30 [schneid]
... scaled comparison, "3h > w"
13:28:31 [sandro]
sandro has joined #owl
13:29:02 [schneid]
... linear comparsion, "3h > xh + s"
13:29:27 [alanr]
action: Bijan to send pointer to nary specification by October 27.
13:29:27 [trackbot]
Created ACTION-234 - Send pointer to nary specification by October 27. [on Bijan Parsia - due 2008-10-30].
13:30:29 [schneid]
alanr: have to decide whether note or rec
13:32:38 [schneid]
cgolbrei: decision of specing or not is not only a question of quality, but also on which level should go in
13:33:48 [schneid]
pfps: proposes to have different vote on letting hooks in or not [fixme]
13:35:13 [schneid]
bparsia: this potential spec does not block any other spec, so would object, if bad review will lead to drop the hooks
13:42:07 [schneid]
ivan: what documents would be involved? want to have a feeling
13:42:30 [schneid]
bmotik: if it is integrated in one of the other documents, then in syntax
13:44:18 [schneid]
alanr: again, will we need the n-aries for rationals?
13:45:44 [schneid]
SubTopic: name of dateType
13:47:07 [sandro]
13:47:09 [wallace]
14:13:15 [Rinke]
Rinke has joined #owl
14:13:52 [sandro]
14:18:33 [schneid]
schneid has joined #owl
14:19:01 [Rinke]
scribenick: Rinke
14:19:50 [Rinke]
Topic: Outstanding XML issues (ISSUE-97)
14:19:58 [Rinke]
Evan: possible agenda change for tomorrow
14:20:11 [Rinke]
Ian: suggested by Bijan to flip the first two sessions of tomorrow
14:20:21 [Rinke]
Ian: any problems with that?
14:20:50 [achilleF]
achilleF has joined #owl
14:21:07 [Rinke]
christine: apologise in advance, as I will be late
14:21:31 [Rinke]
schneid: discussion of future work should be at the end?
14:21:55 [Rinke]
Ian: last week's teleconf we decided that ISSUE-56 should be generalised to a broader future work issue
14:22:46 [Rinke]
Ian: other suggestion 10-10:45 will be the last session, and bump the other sessions
14:23:08 [Rinke]
bijan: that's great
14:23:26 [Rinke]
IanH_: other people ok by that?
14:23:46 [Rinke]
IanH_: moving 10-10:45 session to 16:15
14:24:03 [Rinke]
IanH_: ok, settled.
14:24:10 [Rinke]
IanH_: back to GRDDL discussion
14:24:21 [Rinke]
IanH_: what to say about the GRDDL thing?
14:24:41 [Rinke]
IanH_: we are waiting for an actual XSLT transform to materialise
14:24:55 [Rinke]
alanr: the point is, what level of maturity we need to go to last call
14:25:15 [Rinke]
alanr: until now things have been changing... do we have any dependency on having the GRDDL finished?
14:25:32 [Rinke]
pfps: I suggest we put a pointer to the spec as the GRDDL thing for now
14:25:50 [Rinke]
pfps: do whatever is (minimally) required
14:26:14 [Rinke]
alanr: will changing the content of the GRDDL after last call cause any problems
14:26:43 [Rinke]
sandro: there needs to be a statement saying we may do XSLT in the future, but in the meantime here's the spec that we point to in the meantime
14:27:49 [Rinke]
bijan: this would be a change in design... anything we said about security needs to be adjusted (potential downloading of code)... pointing to the spec would not involve this. It is not merely a change in description.
14:28:11 [Rinke]
bijan: there's a detectable impact on implementations. I
14:28:29 [Rinke]
schneid: for a long time nothing has been discussed?
14:28:42 [Rinke]
bijan: there has been a lot of discussion on the GRDDL group
14:29:00 [Rinke]
schneid: if someone gets up and says I do it... would someone object?
14:29:03 [Rinke]
bijan: yes, I would
14:29:22 [Rinke]
IanH_: last time we pushed this issue off, to see whether the issue would solve itself
14:29:41 [cgolbrei]
cgolbrei has joined #owl
14:29:46 [Rinke]
alanr: it would be some work to do it... but if it's going to be thrown away...
14:29:55 [Rinke]
bijan: but it's still an implementation of the transformation
14:30:38 [Rinke]
alanr: there's a specified place in protocol where GRDDL transforms are found. That's what somebody would be doing it for
14:30:54 [Rinke]
IanH_: what has changed since last time
14:31:56 [Rinke]
ivan: I agree with alan, that the situation today is such that people will not really be willing to do the XSLT because (without going into the discussion on GRDDL interpretation)... the GRDDL users community, regardless of the spec, expects an XSLT transformation in that location
14:32:21 [Rinke]
ivan: people will not invest their time in making an XSLT that 99% of the GRDDL implementations will not find
14:33:10 [Rinke]
bijan: suggest we'll have a decision. Either party can formally object, and we'll have (the chairs?) decide in some way
14:33:31 [Rinke]
sandro: but we cannot give a carte blanche to anyone who promises to build an XSLT
14:34:42 [Rinke]
ivan_: If I start from the principle that someone can come up with an XSLT transformation that passes all the tests, (that we need for OWL/XML) then I would like to see that transformation to be put there and be accessible in that place. But we know that Manchester would be formally agains
14:35:23 [Rinke]
bijan: there's nothing certain about this (e.g. if we have two, limited resources etc.)
14:35:33 [Rinke]
IanH_: suggestions on how to resolve this?
14:35:57 [Rinke]
alanr: we should say what we would consider to be crap and unacceptable, i.e. a call for implementation
14:36:13 [Rinke]
IanH_: would that say anything about if it meets all the ...
14:36:34 [Rinke]
ivan_: somehow we have to be able to say that some implementation that converts owl xml is correct
14:36:44 [Rinke]
ivan_: this xslt transform has to pass all test cases
14:37:08 [Rinke]
IanH_: you didn't say whether you would set a bar 'if it's beyond this level, then we'll definitely use it'
14:37:31 [Rinke]
bijan: I'm a little confused? Did we say we would give out a call for implementations... that's unusual?
14:37:39 [Rinke]
IanH_: no we're considering wording
14:37:45 [Rinke]
ivan: it has happened
14:38:35 [Rinke]
bijan: I have some qualms about that. We've put really high bars on what we included. I'm just weirded out on the special status of this thing wrt. the other parts of the spec.
14:39:00 [Rinke]
bijan: I'm wondering why we're going down this road at all?
14:39:08 [sandro]
14:39:18 [Rinke]
sandro: We asked for volunteers (ACTION-145)... jeremy took this on
14:39:30 [pfps]
pfps has joined #owl
14:39:51 [Rinke]
sandro: Its out of order to say we shoudn't ask for more implementations, because we have established precedent
14:40:11 [Rinke]
sandro: does bijan have a problem with a call for implementations that's neutral wrt. GRDDL
14:40:23 [Rinke]
bijan: no, I'd even contribute to that, potentially
14:40:53 [Rinke]
pfps: I agree entirely. There's no way we can specify a bar for anything to be included
14:41:03 [Rinke]
pfps: the bar would be incredibly high
14:41:25 [sandro]
sandro: cf action-145
14:41:45 [Rinke]
alan: I don't agree. We have the test cases: if it passes the testcases we might deem it ok (that's still to decide though... just to make a point that we can set a bar)
14:42:13 [Rinke]
alanr: we ought to say what we want or don't want, what it is trying to accomplish. If it doesn't do it, then we shouldn't do it.
14:42:13 [sandro]
Sandro: Let's ask for implementations, and encourage people, say it might be for GRDDL, etc, we'll publicize it as an OWL (translator) implementation, etc.
14:42:42 [Rinke]
alanr: We did this for the nary, and I say we should do the same for a GRDDL transformation
14:43:20 [Rinke]
IanH_: that sounds reasonable, but what we said about n-ary... is that if the n-ary thing is spec-standard, then we'll decide on whether we push to include it. If it's not good enough then we chuck it out
14:43:28 [Rinke]
IanH_: are we happy with the same thing for grddl
14:43:48 [Rinke]
alanr: not entirely... if it's not up to standard, it might be a note still
14:44:23 [Rinke]
alanr: we shouldn't try to solve this on procedural grounds.
14:44:51 [Rinke]
alanr: there are some people who think that XSLT for GRDDL transform is bad, and some who don't think it is.
14:45:27 [Rinke]
IanH_: I propose to go forward in exactly the same way as with the nary issue. And do this call, and then decide whether this is up to spec-standard-quality, and move on from there
14:45:58 [sandro]
Ivan: We seem all agreed that it would be good to have the XSLT transformation.
14:46:30 [Rinke]
ivan_: having that XSLT and try to get one is a good case. We agree on that. However, we know in advance that even if this is the bestest XSLT transformation in the whole world, then there will still be objections. This is the difference with the nary issue
14:47:03 [Rinke]
sandro: we will have formal objections either way (no XSLT vs. XSLT)
14:47:13 [FabGandon]
FabGandon has left #owl
14:47:15 [Rinke]
ivan_: other way out is no ref to GRDDL
14:47:23 [Rinke]
(general outings of disagreement)
14:47:32 [Rinke]
bijan: let's not overstate our disagreements here.
14:48:09 [Rinke]
bijan: all this period of time, we already knew that I would have a problem with this... ever since ACTION-145
14:48:24 [Rinke]
bijan: I don't see that the situation has changed?
14:49:01 [Rinke]
schneid: thetwo situations are not the same. nary already a lot of work done, a very weak chance on objection. on the other side, no work done on XSLT, and large chance of objection
14:49:15 [Rinke]
sandro: I don't understand what we're talking about. Nothing has changed since then.
14:49:25 [Rinke]
alanr: there hasn't been a lot done because the spec is not stable
14:49:38 [Rinke]
IanH_: I don't see how this is different from what I've been saying?
14:49:45 [Rinke]
IanH_: we make this call, and see what comes?
14:50:07 [Rinke]
IanH_: what;'s your alternative suggestion? I want to hear! (to alan and ivan)
14:50:45 [Rinke]
ivan_: I don't have a clear alternative suggestion. One suggestion that you could work with (and I would be unhappy) is remove GRDDL. There is no suggestion I would be happy with.
14:51:19 [Rinke]
alanr: grddl pointer to current syntax doc. Can we change this after we've gone for last call?
14:51:31 [Rinke]
alanr: (add the XSLT)
14:51:51 [Rinke]
ivan_: we could add a note that it may change to point to an XSLT
14:52:31 [Rinke]
ivan: there is a precedent. RDF/A made it clear that if there is a proposed recommendation, through the namespace document, an xslt will be included
14:53:09 [Rinke]
alanr: we need to enable whatever we need to enable to make sure that we do not preclude the possibility to add this later
14:53:24 [sandro]
PROPOSED: We keep option-97 (GRDDL XSLT) open through Last Call, not needing to make a decision until PR.
14:53:35 [sandro]
PROPOSED: We keep ISSUE-97 (GRDDL XSLT) open through Last Call, not needing to make a decision until PR.
14:53:36 [Rinke]
IanH_: apparently, if we add such a note we wouldn't need to go through last call again
14:54:22 [Rinke]
IanH_: if no one comes up with one that passes the test cases, then we don't add one (on a remark from schneid on difficulty)
14:54:30 [Rinke]
alanr: should formulate our standards
14:55:02 [sandro]
PROPOSED: We keep ISSUE-97 (GRDDL XSLT) open through Last Call, not needing to make a decision until PR. We'll document this decision as necessary to keep expectations properly set, and we'll keep soliciting a suitable XSLT transform.
14:55:46 [Rinke]
bijan: in the rdfa case they already had agreement that they *wanted* it. But here, we're not all in that situation. I guess I won't be a blocker on this, but I really ahve to ask about myself why I should agree with this (which is not process and not in my interest)
14:56:02 [Rinke]
sandro: last call is looser than CR
14:56:21 [Rinke]
bijan: smaller LC to ask for comments?
14:57:01 [sandro]
PROPOSED: We keep ISSUE-97 (GRDDL XSLT) open through this Last Call, not needing to make a decision until PR. If necessary, we'll do another Last Call for this -- and doing so won't be a reason not to accept the XSLT. We'll document this decision as necessary to keep expectations properly set, and we'll keep soliciting a suitable XSLT transform.
14:57:04 [Rinke]
IanH_: If some xslt comes up, bijan reserves the right to 'tell us' to go back to last call
14:57:44 [IanH_]
14:57:44 [Rinke]
alanr: we shouldnt be blasting over the web, have our requirements in place
14:57:55 [Rinke]
alanr: first
14:58:07 [bijan]
bijan has joined #owl
14:58:14 [sandro]
PROPOSED: We keep ISSUE-97 (GRDDL XSLT) open through this Last Call, not needing to make a decision until PR. If necessary, we'll do another Last Call for this -- and doing so won't be a reason not to accept the XSLT. We'll document this decision as necessary to keep expectations properly set, and we'll keep soliciting a suitable XSLT transform.
14:58:45 [Rinke]
IanH_: apart from the separate decision on what would constitute a satisfactory/suitable XSLT should do
14:59:02 [Rinke]
IanH_: we don't need that in this resolution
14:59:25 [Rinke]
pfps: the reason for requiring a second last call is that we put a pointer?
14:59:32 [Rinke]
bijan: no, it would constitute a change
14:59:46 [Rinke]
sandro: that's this line about sufficient documentation
15:00:45 [Rinke]
pfps: I would like to have us decide now that our at least temporary grddl transform says "do what the spec says"
15:02:05 [Rinke]
ivan: there is a disagreement between bijan and the WG members of the group, that yes it is allowed legally to have something in there which is not an XSLT, but the whole community expects an XSLT there. It might be correct, but nobody expects it
15:02:20 [Rinke]
pfps: it's a recommendation. We're doing the right thing according the rec, case closed
15:02:29 [Rinke]
ivan: we're doing it for the community
15:03:05 [Rinke]
bijan: I do agree with you on this
15:03:29 [Rinke]
IanH_: there's a simple thing to decide on this. Leave the GRDDL empty, or point to the spec?
15:04:04 [Rinke]
pfps: if we're doing this for the community, as opposed to a de jure requirement. If it's not a requirement, then don't do it!
15:04:10 [Rinke]
pfps: let's just put nothing there
15:05:25 [Rinke]
alanr: leaving a pointer, is not what the community expects. In terms of being friendly to the community that's not the best idea. I won't object to it... but it is really a statement by some members of the OWL WG, and does not communicate what it was supposed to do.
15:05:45 [Rinke]
alanr: put nothing there in the meantime
15:05:59 [Rinke]
sandro: I didn't hear any objection to my proposal right?
15:06:13 [sandro]
PROPOSED: We keep ISSUE-97 (GRDDL XSLT) open through this Last Call, not needing to make a decision until PR. If necessary, we'll do another Last Call for this -- and doing so won't be a reason not to accept the XSLT. We'll document this decision as necessary to keep expectations properly set, and we'll keep soliciting a suitable XSLT transform.
15:06:14 [Rinke]
sandro: we could deal with my proposal first, and then continue the discussion
15:06:21 [sandro]
15:06:23 [Rinke]
pfps: I'm happy to vote on the proposal
15:06:31 [ivan_]
15:08:30 [Rinke]
bijan: the proposal as it stands has a presumption that I won't go out and badmouth it. It might be good to just put our cards on the table and go ... at this moment I abstain, as I am not sure that this way of going is moving us anything closer to consensus
15:08:35 [pfps]
15:08:35 [Rinke]
IanH_: we already accepted that
15:08:39 [bijan]
15:08:47 [Rinke]
sandro: there's no phrasing that would improve it?
15:08:47 [alanr]
15:08:49 [Rinke]
bijan: no
15:08:52 [wallace]
15:08:54 [Zhe]
15:08:54 [achilleF]
15:08:56 [ivan_]
15:08:57 [MarkusK_]
15:09:00 [bernardo]
15:09:02 [Rinke]
15:09:09 [schneid]
15:09:17 [cgolbrei]
0 (uvsq)
15:09:34 [IanH_]
15:09:54 [Rinke]
bijan: there doesn't seem to be strong support on this
15:10:18 [Rinke]
IanH_: we've got weak acceptance for this, but no objection
15:10:27 [Rinke]
alan: so this proposal is resolved
15:10:47 [sandro]
RESOLVED: We keep ISSUE-97 (GRDDL XSLT) open through this Last Call, not needing to make a decision until PR. If necessary, we'll do another Last Call for this -- and doing so won't be a reason not to accept the XSLT. We'll document this decision as necessary to keep expectations properly set, and we'll keep soliciting a suitable XSLT transform.
15:11:27 [Rinke]
bijan: I agree in general that spec worship is bad. I prefer in general to break spec-backwardscompatibility.......
15:11:29 [msmith1]
msmith1 has joined #owl
15:11:48 [Rinke]
bijan: if you appeal to what the grddl community wants, then that's not convincing.
15:12:09 [Rinke]
sandro: it's too bad the GRDDL wasn't more clear on this, allowing you to be more clear in your objection
15:12:53 [Rinke]
schneid: general question about bugs. For all recommendations there is an erratum, how do these bugfixes get into the errata. How is it created?
15:13:06 [Rinke]
ivan_: it is the responsibility of the WG to set up a mechanism.
15:13:16 [Rinke]
ivan_: we took into account the errata for OWL 1
15:13:51 [Rinke]
schneid: could there not be a way out if bijan et al. fixed the loophole?
15:14:20 [Rinke]
bijan: I would have objected. The loophole is the right way to go
15:14:26 [Rinke]
pfps: the loophole is the right way to do
15:14:30 [Rinke]
15:14:47 [sandro]
sandro: I'd like to understand why you think it's the right way to go.
15:15:02 [Rinke]
alan: is it a good idea to have everyone who has an idea on how to guarantee quality on this transform put this on email
15:15:21 [Rinke]
alan: I don't know what the bar is, I'd like to have some idea...
15:15:52 [sandro]
Peter: An automatic process that would produce the XSLT from our documents.
15:15:55 [Rinke]
pfps: a provably correct automated process that would derive an XSLT transform from our documents.
15:16:00 [sandro]
Sandro: Yes. That would be great!
15:16:15 [Rinke]
bijan: yes, that would be a first step
15:17:18 [Rinke]
bijan: that would be nice... other things that would be nice is if it were software-readable. Performance is an issue... it has to have a very clear forward errata process (my early objection was that the initial version would be a default standard)
15:17:34 [Rinke]
alan: doesn't this hold for the schema document as well?
15:18:43 [Rinke]
bijan: we need an errata process in general. My point is that it is much less likely that an XML schema is downloaded automatically... and it does not produce any meaning. An XSLT really changes something that I put out... xml schema does checking, and I would need to explicitly invoke this
15:19:16 [Rinke]
sandro: you say 'typical', but schema can be used to generate parse tree (?)
15:19:27 [Rinke]
bijan: you mean the schema is at the location of namespace?
15:19:38 [Rinke]
sandro: yes, at least dereferencable
15:19:46 [Rinke]
bijan: I'm not sure I would support that
15:19:51 [alanr]
alanr has joined #owl
15:19:55 [Rinke]
alan: issues are very helpful! Location of schema etc.
15:20:15 [bijan]
Also, in general, scarily, XML Schema is easier to understand
15:20:21 [bijan]
At least, have confidence that it is correct
15:20:23 [bijan]
For me at least
15:20:28 [Rinke]
IanH_: we have some guidelines for what's good quality. No procedure for asking for the implementation
15:20:55 [Rinke]
alanr: do we need to decide now to as a WG send out a notice to some list?
15:21:00 [Rinke]
sando: I assume it's editorial
15:21:10 [Rinke]
IanH_: are we done with this issue?
15:21:26 [Rinke]
alanr: the only thing... where we going to point to the specification?
15:21:42 [Rinke]
IanH_: point or do nothing
15:21:54 [Rinke]
sandro, bijan, ivan, alan, pfps: not do anything
15:22:14 [Rinke]
bijan: mention of grddl in the charter means that we still...
15:22:24 [Rinke]
sadro: we need to specify how to deal with grddl
15:22:36 [Rinke]
ivan: if there's a deviation from the charter, we need to document htis
15:22:46 [Rinke]
IanH_: that's for later
15:23:06 [sandro]
<sandro> PROPOSED: We won't use GRDDL until we figure out what we're doing with GRDDL. By charter, we'll have to figure out eventually what we're doing with GRDDL. At the namespace document was can refer to this issue.\
15:23:10 [Rinke]
alanr: if it turns out that the strategy is to read the spec by a program i.e. via css... class markup
15:23:15 [sandro]
PROPOSED: We won't use GRDDL until we figure out what we're doing with GRDDL. By charter, we'll have to figure out eventually what we're doing with GRDDL. At the namespace document was can refer to this issue.\
15:24:34 [Rinke]
pfps: the resolution that we might end up with is that we don't use grddl....
15:24:56 [sandro]
PROPOSED: We won't use GRDDL until we figure out what we're doing with GRDDL. By charter, we'll have to figure out eventually what we're doing with GRDDL. At the namespace document, for now, was can refer to this issue.
15:25:03 [sandro]
PROPOSED: We won't use GRDDL until we figure out what we're doing with GRDDL. By charter, we'll have to figure out eventually what we're doing with GRDDL. At the namespace document, for now, we can refer to this issue.
15:25:31 [Rinke]
pfps: I would prefer to say put nothing on the namespace document at all, and put our current lack of consensus elsewhere
15:26:28 [Rinke]
sando: if there's nothing there, grddl-people who look at it will say oh my god, you're not doing grddl
15:26:37 [sandro]
15:26:43 [ivan_]
15:26:43 [bijan]
15:26:46 [pfps]
15:26:46 [IanH_]
15:26:49 [alanr]
15:26:51 [Rinke]
pfps: ok, if it has to be there (Because that's where they'll look) that's ok
15:26:51 [Rinke]
15:26:53 [achilleF]
15:26:54 [MarkusK_]
15:26:54 [bernardo]
15:27:30 [Rinke]
IanH_: that seems to be resolved, more or less unanymously
15:27:31 [sandro]
RESOLVED: We won't use GRDDL until we figure out what we're doing with GRDDL. By charter, we'll have to figure out eventually what we're doing with GRDDL. At the namespace document, for now, we can refer to this issue.
15:27:52 [Rinke]
ivan_: there was one discussion we had today that was postponed... should it be recorded as issue perhaps...
15:28:02 [Rinke]
ivan_: which are the normative serialisations of owl
15:28:14 [Rinke]
alanr: yes, open an issue
15:28:27 [Rinke]
ivan_: rdf vs. owl xml
15:28:35 [sandro]
ISSUE: Which serialization of OWL are suitable as normative exchange syntaxed?
15:28:36 [trackbot]
Created ISSUE-150 - Which serialization of OWL are suitable as normative exchange syntaxed? ; please complete additional details at .
15:29:56 [sandro]
Bijan: I think some folks outside the group may object to any normative exchange syntax other than RDF/XML
15:30:00 [Rinke]
bijan: there might be some tactical issues here
15:30:31 [sandro]
Sandro: I think GRDDL+XSLT might be enough.
15:30:38 [Rinke]
pfps: rdf/xml is the default syntax for xml, but content providers might negotiate the exchange of other syntaxes as well
15:30:53 [Rinke]
pfps: you have to be prepared to serve RDF/XML for any OWL 2 ontology
15:31:20 [Rinke]
pfps: if you exchange in the FS, you use this mimetype and syntax (normatively). If you ask for XML syntax the same....
15:31:42 [sandro]
PROPOSED: Close issue-150 saying RDF/XML is the normative exchange syntax for OWL 2. Systems should transmit other OWL 2 serializations only when there has been prior arrangement to use the alternative serialization (eg by HTTP Content Negotiation).
15:31:46 [Rinke]
pfps: if you are a consumer, and you say you only want the XML syntax, the provider may say 'sorry I don't do that'
15:32:01 [Rinke]
pfps: but if you ask for RDF/XML you have to serve it
15:32:21 [Rinke]
bijan: MUST support RDF/XML, MAY or SHOULD the others
15:32:27 [Rinke]
pfps: I like SHOULD
15:33:04 [Rinke]
sandro: there's some toughness here, but the spirit ...
15:33:17 [Rinke]
ivan: at some point in time there has to be some wordsmithing, but the intent is ok
15:34:08 [sandro]
pfps: The idea is: you HAVE TO serve RDF/XML, but you may provide others.
15:34:37 [Rinke]
sandro: may even specify weights
15:34:55 [Rinke]
ivan: you can rank what syntaxes you prefer
15:35:01 [Rinke]
pfps: may even be completely agnostic
15:35:08 [Rinke]
15:35:30 [Rinke]
alanr: do we want to say something on conformance of tools to take this.
15:35:41 [Rinke]
ivan_: every tool has to take RDF/XML
15:36:24 [Rinke]
pfps: not quite. I could be an editor (damaged P4, that is limited to preserve roundtripping), but could still do something
15:36:54 [Rinke]
bijan: I'm ok with this in general. If I write a tool that only consumes OWL XML... do we care? I mean, who cares? there's only so much influence we can have here
15:37:26 [Rinke]
bijan: are we commited to having an errata later on if the facts on the ground change (e.g.Turtle becomes the defacto standard)
15:37:42 [Rinke]
bijan: the more conformance you enforce that does not make sense, the weaker your power
15:39:04 [Rinke]
alan: does normative RDF/XML mean that every tool must be able to exchange this syntax to be conformant?
15:39:06 [Rinke]
ivan: yes
15:40:06 [sandro]
PROPOSED: Close issue-150 saying conformant OWL reasoners MUST accept rdf/xml. They MAY accept other serializations. Systems publishing OWL2 should publish it in RDF/XML; they MAY also publish in other serialzations, but should only send it by prior arrangement (eg content-negotiation).
15:40:08 [Rinke]
bijan: I don't object, but don't think it's going to do anything useful. Any tool that is doing anything useful will do RDF/XML.... but my student won't do this as it's hard to parse
15:40:24 [Rinke]
bijan: what vendor's mind is going to be affected by this decision
15:40:48 [Rinke]
pfps: the rason for saying this is to tryo and diffuse the potential laying down on the ground and scream
15:40:53 [Rinke]
15:40:57 [sandro]
peter: the reason for this is to diffuse the lay-down-on-the-ground-and-scream reaction.
15:40:57 [Rinke]
15:41:21 [Rinke]
schneid: we had normativity for RDF/XML in OWL 1
15:41:28 [Rinke]
schneid: would it be possible not to have this in OWL 2
15:41:40 [Rinke]
IanH_: It doesn't matter if we decide to do it...
15:41:52 [Rinke]
bijan: If you don't want to make it non-normative, then your question is moot
15:42:08 [Rinke]
IanH_: does the proposal as it stands capture what it should
15:42:39 [Rinke]
wallace: it seems superfluous. It's fine, but I don't understand why this would be needed.
15:43:07 [Rinke]
pfps: I believe it is useful because there is a community that very much want it to be 'should not' produce other syntaxes
15:43:13 [sandro]
peter: This is a community that wanted "SHOULD NOT" instead of "MAY".
15:43:52 [Rinke]
pfps: original was MUST be able instead of MAY
15:44:43 [Rinke]
alanr: it is useful to have a default... predictability is nice
15:45:27 [Rinke]
pfps: this is not the way it should be. Suppose I'm a repository of P4 documents, mastered in DL style. And bijan is a pellet reasoner, Under the should, if bijan doesn't care, I take my FS turn it into RDF/XML and bijan does it the other way around.
15:45:58 [Rinke]
alan: but if you have a random user, then default is easier (because you encounter all these different syntaxes)
15:46:20 [Rinke]
bijan: if you build a crawler, you can specify what format you want to retrieve
15:47:26 [Rinke]
ivan_: we may have the shortest-lived issue ever
15:47:37 [Rinke]
ivan_: it wasn't even open yet
15:47:39 [sandro]
PROPOSED: Close issue-150 saying conformant OWL reasoners MUST accept RDF/XML. They MAY accept other serializations. Systems publishing OWL2 MUST publish it in RDF/XML if asked; they MAY also publish in other serialzations, but only by prior arrangement (eg content-negotiation).
15:47:51 [Rinke]
15:48:14 [pfps]
PROPOSED: Close issue-150 saying conformant OWL reasoners MUST accept RDF/XML. They MAY accept other serializations. Systems publishing OWL2 MUST publish it in RDF/XML if asked; they MAY also publish in other serialzations
15:48:47 [Rinke]
15:49:42 [sandro]
<pfps> PROPOSED: Close issue-150 saying conformant OWL reasoners MUST accept RDF/XML. They MAY accept other serializations. Systems publishing OWL2 MUST publish it in RDF/XML if asked (eg with HTTP content-negotiation); they MAY also publish in other serializations.
15:50:12 [sandro]
PROPOSED: Close issue-150 saying conformant OWL reasoners MUST accept RDF/XML. They MAY accept other serializations. Systems publishing OWL2 MUST publish it in RDF/XML if asked (eg with HTTP content-negotiation); they MAY also publish in other serializations.
15:50:15 [sandro]
15:50:17 [Rinke]
15:50:18 [MarkusK_]
15:50:20 [wallace]
15:50:20 [alanr]
15:50:20 [ivan_]
15:50:21 [pfps]
15:50:22 [bernardo]
15:50:22 [achilleF]
15:50:24 [achilleF]
15:50:24 [IanH_]
15:50:28 [bijan]
15:50:30 [cgolbrei]
15:50:34 [sandro]
RESOLVED: Close issue-150 saying conformant OWL reasoners MUST accept RDF/XML. They MAY accept other serializations. Systems publishing OWL2 MUST publish it in RDF/XML if asked (eg with HTTP content-negotiation); they MAY also publish in other serializations.
15:51:02 [Rinke]
IanH_: I think we're getting very close to the end of the session
15:51:07 [Rinke]
ivan_: we settled everything?
15:51:11 [Rinke]
pfps: no, not xsd datetype
15:51:12 [sandro]
RRSAgent, pointer?
15:51:12 [RRSAgent]
15:51:22 [Rinke]
IanH_: can we do something on that in the 10 minutes left?
15:51:32 [Rinke]
pfps: I'll meet henry tomorrow to talk on this
15:51:51 [Rinke]
ivan_: one question on the manchester syntax
15:52:06 [Rinke]
ivan_: can I put a comment in the manchester syntax?
15:52:47 [Rinke]
pfps: there was a discussion in webont that annotations should be comments, this was not accepted.
15:53:25 [Rinke]
ivan: in RDF/XML I can do XML comments, in Turtle I can use the hash
15:53:31 [Rinke]
pfps: perfectly happy to have
15:53:46 [Rinke]
bmotik: that should be easily changed
15:54:09 [Rinke]
bijan: caveats should be explicit
15:54:18 [Rinke]
pfps: comments may be stripped: this should be explicitly mentioned
15:54:54 [Rinke]
bijan: the OWL-S group used xml comments and had trouble using that. So a statement that warns people is useful
15:55:01 [Rinke]
pfps: I think it's a good idea to add
15:55:32 [Rinke]
IanH_: editorial for peter and bmotik
15:55:54 [Rinke]
"whatever turtle has:"
15:56:53 [Rinke]
IanH_: are we done?
15:57:05 [Rinke]
sandro: comment on scribing. I've been making cleanups as we went along
15:57:24 [Rinke]
sandro: in the wiki... topics, subtopics etc.... I wrote a script that did that realtime
15:57:38 [Rinke]
IanH_: talking of scribing... about tomorrow
15:58:06 [Rinke]
boris: I could scribe the session after lunch
15:58:17 [Rinke]
IanH_: volunteers for the first session
15:58:22 [Rinke]
wallace: I will do that
15:58:28 [Rinke]
boris: I will go before lunch
15:59:02 [Rinke]
achilleF: first session after lunch
15:59:19 [Rinke]
ivan: roadmap
15:59:56 [sandro]
15:59:59 [sandro]
talking about dinner.
16:00:52 [sandro]
Meet at hotel reception desk at 7pm
16:01:08 [sandro]
RRSAgent, pointer?
16:01:08 [RRSAgent]
16:01:49 [IanH_]
zakim, who is here?
16:01:49 [Zakim]
On the phone I see Riviera_B, Zhe
16:01:50 [Zakim]
On IRC I see alanr, msmith, bijan, pfps, cgolbrei, achilleF, schneid, sandro, bernardo, MarkusK_, wallace, IanH_, Zhe, RRSAgent, Zakim, trackbot
16:02:00 [IanH_]
Zhe -- are you still there?
16:02:06 [Zhe]
16:02:24 [Zhe]
I am not on the phone right now
16:02:51 [IanH_]
Will you be around first thing tomorrow morning for the discussion on issue 144?
16:03:09 [Zhe]
you mean 3am my local time :)
16:03:23 [bijan]
Yeah, Elisa stayed up!
16:03:28 [IanH_]
Yes -- and I wouldn't even ask except that you were here at that time today :-)
16:03:31 [bijan]
You're not going to let her punk you, eh?
16:03:43 [Zhe]
I will try
16:04:14 [IanH_]
We could reschedule if it would help?
16:04:56 [Zhe]
make it after 3:15am (EST) can help
16:05:34 [IanH_]
So this would be after 9:15 am French time?
16:05:45 [Zhe]
i think so.
16:05:57 [IanH_]
So this is middle of the night your time?
16:06:03 [Zhe]
16:06:09 [IanH_]
16:06:18 [Zhe]
16:06:22 [IanH_]
We admire your devotion!
16:06:46 [Zhe]
I will be half asleep ...
16:06:49 [IanH_]
Or do you have to get up at this time to feed the baby?
16:06:50 [Zhe]
16:07:19 [Zhe]
luckily, I don't have to... my wife takes that duty
16:07:31 [IanH_]
Nice :-)
16:08:28 [Zhe]
have a good dinner/night
16:08:41 [IanH_]
Thanks -- and the same to you!
16:08:44 [Zakim]
16:16:26 [alanr]
16:16:29 [alanr]
zakim, pointer
16:16:29 [Zakim]
I don't understand 'pointer', alanr
16:16:33 [alanr]
rrsagent, pointer
16:16:33 [RRSAgent]
16:35:00 [Zakim]
disconnecting the lone participant, Zhe, in SW_OWL(F2F)2:30AM
16:35:03 [Zakim]
SW_OWL(F2F)2:30AM has ended
16:35:04 [Zakim]
Attendees were Elisa_Kendall, Riviera_B, Riviera_B.a, Zhe
16:44:29 [IanH]
IanH has joined #owl
18:21:22 [Zakim]
Zakim has left #owl
20:25:03 [IanH_]
IanH_ has joined #owl
20:26:22 [Rinke]
Rinke has joined #owl