ADMIN: Minutes Dec 12

=====

Executive Summary:

DECISIONS:

RESOLVED:
- Open and Postpone Justification issue per Dan Connolly's email.
- Accept items 1, 2, and 5 of Peter Patel-Scheider's proposed closure
  for Issue 5.8 - datatypes.
- Close Issue 5.13 by advising users in the documents that they can use:
  application/xml, application/rdf+xml, application/owl+xml.  
- Close Issue 5.23 by resolving to not add hasValue to OWL Lite.


NEW ACTIONS:
- Mike Smith to add Postponed Justification issue to issue list.
- Jeremy to create text to use in place of PPS item 3 in resolution to
Issue 5.8 on datatypes.
- Dan C to to communicate with XML schema group about URIs for
XML datatypes.
- Jonathon Borden to update media types document and pass to M Dean
for use in editing Reference Document.
- Dan C take media type registration request to IETF.
- Ian to write up an explanation of known characteristics
for decision procedures for OWL Lite and OWL DL.

-----------------

More detail:

IRC:
http://www.w3.org/2002/12/12-webont-irc

AGENDA:
http://lists.w3.org/Archives/Public/www-webont-wg/2002Dec/0106.html


1) ADMIN (15 min)

1.1 Role Call

Dan Connolly, Marwan Sabbouh, Evan Wallace, Peter Patel-Schneider,
Mike Smith, Ian Horrocks, Mike Dean, Tim Finin, Ruediger Klein, Jeremy
Carroll, Jeff Heflin, Masahiro Hori, Jim Hendler, Ziv Hellman, Herman
ter Horst, Deborah McGuinness, Jos De Roo, Pat Hayes, Frank van Harmelen,
Jean-Francois Baget (irc), Nicholas Gibbins

Late: Jonathan Borden

Regrets: Guus Schreiber, Lynn Stein, Jonathan Dale, Larry Eshelman,
John Stanton

1.2 Approval of Minutes of Dec 5 call

PROPOSED to accept the following as a true record of the Dec 5
telecon:
http://www.w3.org/2002/12/05-webont-irc


Dan prefered Jeremy's notes as sent in message 
http://lists.w3.org/Archives/Public/www-webont-wg/2002Dec/0114.html
subject to revision by Jeremy (link to previous meeting, and M Smith 
ammendment etc).

ACCEPTED, see final version in subsequent message:
http://lists.w3.org/Archives/Public/www-webont-wg/2002Dec/0172.html


1.3 Agenda Amendments

Dan Connolly's Justifications issue, as described in his email 
http://lists.w3.org/Archives/Public/www-webont-wg/2002Dec/0110.html


1.4   Telecon Schedule

Next telecon: December 19
Scribe:   Ian Horrocks, if present

Question was raised: should we cancel Dec 26 telecon?
No one objected.  Dec 26 meeting CANCELLED!

2 Jan 03 meeting NOT cancelled.
Dan Connolly notes potential regrets for that meeting.


1.5  F2F Meeting Manchester

Registration page:
http://cgi.w3.org/Register/selectUser.pl?_w3c_meetingName=WebontManchester

Local arrangements page:
http://wonderweb.semanticweb.org/mcr-f2f.shtml

Short discussion of plan for f2f: Hendler

Jim Hendler reviewed plans for a meeting focus on implementation and
interoperability experience with OWL.  In addition, detailed working
draft editorial inspections will be undertaken.

Mike Smith started a discussion on the preferred anchor pattern to use
in preparing WDs for this.  Advice was: provide anchors for all
language elements, and use same case for anchor strings as that used in
wd text for human consumption.

Insertiong of anchors and other preparations of the Working Drafts
should be completed a week before the ftf (2 January 2003).

1.6 Report on Web Ontology Working Group Extension
AC Vote period closed.  24 votes cast - 23 approve, 1 abstain
Information passed on to W3M for decision
If/when extension granted, members not yet re-enrolled will be informed


1.7 Webont schedule

FTI: Current schedule:
- LC to start directly after Man ftf (around Jan 15)
- LC period 3-6 weeks
- Man ftf and LC period to be used by WG to gather implementation
    experiences
- odds are low that a WebOnt ftf is required in March at Tech Plenary

Jeremy Carroll:

  Colleagues at HP believe Jeremy was too negative in his concerns
  about schedule as expressed in email:
  http://lists.w3.org/Archives/Public/www-webont-wg/2002Dec/0048.html.
  He suggests a longer Last Call period (e.g. six weeks) may be
  sufficient to address these concerns.

Jim Hendler:

  Chairs leaning toward lengthening Last Call period.  May go to last
  call after Jan Face-to-face.

Dan Connolly:

  In light of above, do we need a face-to-face in March (in conjunction
  with the Technical plenary)?

Chairs and W3C contact to discuss this further off line.

2. ISSUES (60 min)

Link to issues list:
http://www.w3.org/2001/sw/WebOnt/webont-issues.html

2.1  Pending updates:

ISSUE 5.14 Versioning Closed Dec. 05
closing text at:
http://lists.w3.org/Archives/Public/www-webont-wg/2002Dec/0075.html
(Versioning needs to go into the reference and maybe Guide documents)
(Issues List also needs to reflect this change -EKW as scribe)

2.2 ACTIONS wrt. resolved issues

ACTION Chris Welty will work on getting "The meaning of
owl:ontology" explained better in Guide, other editors will see
if their documents need changing (not obvious they do).
[Continued]

ACTION Jeremy to generate test Cases for 5.5. List Syntax or Semantics
[Continued]

ACTION Jeff Heflin will produce test cases for owl:imports
See also message from Jeremy Carroll:
http://lists.w3.org/Archives/Public/www-webont-wg/2002Nov/0228.html
[Done]

ACTION DanC to provide wording
[Concerns Issue 5.9 - malformed D+O restrictions (closed)
Dan would like to amend the proposal with some clear instructions to the
guide that says "Don't do that".  Dan was actioned to write up a few
sentences for the guide.]
[Done - see email
http://lists.w3.org/Archives/Public/www-webont-wg/2002Dec/0127.html]

ACTION: Deb/Change feature doc in accordance with 5.19 resolution
(using MikeS's text)
[Continued]

ACTION: MikeDean/update reference appropriately for three
sublanguages
[Continued]

ACTION: Deb/Change features for three sublanguages
[Continued]


3.2 OPEN issues

New issue: Justification 

RESOLVED by consensus to open and postpone this issue per Dan's email.
http://lists.w3.org/Archives/Public/www-webont-wg/2002Dec/0110.html
NEW ACTION: Mike Smith to add this to issue list

[We have five open issues left]

** Issue 5.8 - datatypes

Peter's proposal
http://lists.w3.org/Archives/Public/www-webont-wg/2002Nov/0265.html

Pat Hayes - 
  Does bullet five prevent referenced literals?

other discussion.  

Hendler put question on items 1, 2 and 5 of Peters message.  

RESOLVED 
 objections: none
 abstentions: Mike Dean, Jeff Heflin


Discussion of problems with 3:

Prior email comments on this from Jeremy:
http://lists.w3.org/Archives/Public/www-webont-wg/2002Nov/0278.html
http://lists.w3.org/Archives/Public/www-webont-wg/2002Dec/0115.html

Jeremy summarized these comments, identifying a few XML Schema
built-in non-list types which present problems for OWL (and RDF for
that matter). [Had trouble hearing which, although I assume they were
those discussed in email. -EKW as scribe] 

Discussion ensued, such as:
 Dan: QNAMEs do not behave as RDF datatypes.
      
 Jeremy:  We should specify XML schema datatypes that we will
 support, but note that there are ones that we won't support.

 Dan: It is worthwhile to include a note explaining why this was
      not the complete XML list.

NEW ACTION Jeremy:
  Write up suggested text to address item 3.

see html attached to a subsequent email from Jeremy:
http://lists.w3.org/Archives/Public/www-webont-wg/2002Dec/att-0174/01-fo.HTML


Item 4:

Chair's decision on part 4 of this proposal (out of scope)
http://lists.w3.org/Archives/Public/www-webont-wg/2002Dec/0104.html

Discussion:
Peter would hate not to use a solution that solves 99% of
this problem.

Dan points out that what we need to do is to include the XML Schema
group in our solution group.  This potentially creates a schedule
risk in order to "do the right thing."

Jeremy has an action from last week to draft a message for wg
consideration as a message to tag about item 4.

NEW ACTION Dan - to communicate with XML schema group about URIs for
XML datatypes.

** Issue 5.13 - Internet media type for OWL

ACTION: chairs, to ask Jonathan Borden re: wording to close 5.13 - DONE

Proposal to close  (As per last week's request for changes):
http://lists.w3.org/Archives/Public/www-webont-wg/2002Dec/0057.html

Discussion:
Jonathan changed entailment parameter to a URI per suggestion

Dan skeptical that a test could be defined to demonstrate use
of this parameter.

Mike Dean: doesn't know how to tag his current data with this 
parameter.

Jonathon: How do you tell what entailments are licensed for an OWL
document without this.  

general: some discussion about inferring from vocabulary.

Frank van Harmelan : It's not just the vocabulary that suggests intention,
it is also the way the vocabulary is put together.

Hendler: We some way to declare intent for our usage of OWL.

Dan convinced that it is worth having mime types for OWL Lite and 
OWL DL (OWL Full is default entailment).

PHayes: use semantic extension vocabulary.

Dan's proposal: 4 mime types (same as 30 Oct mail?)
   xml
   rdf+xml
   owl-lite
   owl-dl
???owl-full???

Dan explanation of http protocol  --client states preference.

Mike Dean best practice: anything that uses OWL namespace should use
 owl mime type.

Jim Hendler - proposal
  advise users in the documents that they can use:
application/xml, application/rdf+xml, application/owl+xml

Straw poll - Pat Hayes objected
  Pat hayes then suggests changing app/owl+xml to app/owl+rdf+xml

Reference document or Guide would be the target for this advice.

Called question as proposed by JH above, not with Pat's change.
RESOLVED
 objections: none
 abstentions: Dan Connolly, Frank van Harmelen, Jeremy Carroll, Pat
 Hayes, Jos De Roo

NEW ACTION: 
 Jonathon Borden to update document and pass to M Dean for use
in editing Reference Document.

NEW ACTION: 
 Dan C take media type registration request to IETF 


** Issue 15.23: add hasValue to Owl Lite

[sound and completeness discussion]

A great deal of email discussion took place prior to this meeting on
the subject of how the inclusion of hasValue in OWL Lite would affect
the implementation of OWL Lite reasoners.  Two of the strongest voices
on each side of this debate: Ian Horrocks and Jim Hendler continued
this discussion by phone.  This revealed some details with regard to 
prior implementations that may have not been fully understood.
http://lists.w3.org/Archives/Public/www-webont-wg/2002Dec/0160.html
Describes this.  This sets the background for the discussion on this
at the telecon.

Ian asserts that - There are known decision procedures for OWL Lite
that are sound, complete and terminating.  Implementations exist for
this.  Addition of hasValue would break this property.

Note that because OWL DL includes both hasValue and oneOf, decision
procedures for OWL DL reasoning are not _currently known_ to have these
properties.

NEW ACTION Ian Horrocks to write an explanation of this situation

Given this information about known repercussions of including hasValue
in OWL language subsets, the chair took a straw pole concerning
leaving hasValue out of OWL Lite.  No one responded that they could
not live with such an action.

The question was put:
  Close Issue 5.23 by resolving to not add hasValue to OWL Lite.
RESOLVED.
 objections: none
 abstentions: Dan Connolly, Jos De Roo, Jonathan Borden, Jeremy Carroll,
 Deb McGuinness, Mike Dean, Pat Hayes, Jeff Heflin, Jim Hendler?

Discussion about the OWL DL feature set and implementation of OWL DL
reasoners was not entertained during the telecon, but was welcomed at
a future time (in subsequent email or when scheduled for subsequent
telecons).

Received on Tuesday, 17 December 2002 17:31:57 UTC