ADMIN: Minutes of 2/20/03 Telecon

WEB ONTOLOGY WORKING GROUP TELECON
February 20, 2003

Chair: Jim Hendler
Scribe: Leo Obrst

Chat Log: http://www.w3.org/2003/02/20-webont-irc#T19-00-40

Quotes are approximate; apologies for any misattributions.


SUMMARY OF NEW ACTIONS, RESOLUTIONS:

1.2 Approval of Minutes of last telecon
NEW ACTION: Jeff will update and post revision to list.

1.5  Tech plenary
NEW ACTION, Chairs: Agenda needed for f2f plenary.
NEW ACTION, Chairs: speaker phone at Tech plenary on Tuesday

3.4 Relation between RDFS and OWL Lite
NEW ACTION: Jeremy to update his discussion of this issue, give quick
comments.
NEW ACTION: Guus to review, add to the Reference document as Appendix.
NEW ACTION: Deb to add a paragraph to Overview doc.
NEW ACTION: Jim, Pat to review

2. RESPONSE TO RDFCore LC
ACTION Guus summary review of RDF Schema doc as proposal for the WG
PROPOSAL: this is our final version of the comments
PROPOSAL passes, resolved. THIS ACTION ITEM CLOSED.
NEW ACTION: Jim to send comments to RDF CORE

3.2 OWL DL syntax
NEW ACTION: Jim to open this as an issue.
NEW ACTION: Mike Smith to put (message Jan 04089) in issues document.
NEW ACTION: Ian, Jeremy, Peter, Pat, Sean, Guus, Jos will work as task
force on
discussion and return in 1 week with progress. Probably can't resolve
this next week.

4. DOCUMENT STATUS (10-15 min.)
NEW ACTION: Evan and Mike Dean to take MIME type and produce MIME type
doc.


DETAIL:

1) ADMIN (10-15 min)

1.1 Role Call

JimH, PatH, DanC, Pfps, MikeSmith, Jeremy, Leo, Ian_Horrocks, Nick,
Evan_Wallace, Mike_Dean, John_Stanton, DebM, JeffH, Guus, Herman,
TimFinin, JeromeE, Raphael, Massimo, SeanB, Marwan, Jos

Regrets: Klein, Baget, Thompson, alSafadi, Dale
See list archive for last-minute regrets

1.2 Approval of Minutes of last telecon

PROPOSED to accept the following as a true record of Feb 13
http://lists.w3.org/Archives/Public/www-webont-wg/2003Feb/0193.html

Jeff: one minor change, Jonathan Borden was present.
NEW ACTION: Jeff will update and post to list revision.

1.3 Agenda Amendments
Agenda item 3.4 moved to initial position in the Agenda.

1.4 Telecon Schedule

Next telecon: February 27
Chair: Jim Hendler
Scribe: Jérôme Euzenat

1.5  Tech plenary

Tech plenary page:
http://www.w3.org/2002/10/allgroupoverview.html

- Mar 4: editor's meeting on Tuesday

Deadline over for AV requests, same for Mar 6-7.

- Mar 6-7: Initial Tech Plenary Semantic Web Architecture page -
http://www.w3.org/2001/sw/meetings/tech-200303/

Discussion:
45 folks registered.
Syntactic issues: RDF inside of HTML
Jeremy Carroll will run the social meaning session.

NEW ACTION, Chairs: Agenda needed for f2f plenary.

NEW ACTION Chairs: speaker phone at Tech plenary on Tuesday

3.4 Relation between RDFS and OWL Lite
(THIS MOVED UP TO FIRST AGENDA ITEM)

Context: see message by Hendler:
http://lists.w3.org/Archives/Public/www-webont-wg/2003Feb/0130.html

See also prior discussion in October:
http://lists.w3.org/Archives/Public/www-webont-wg/2002Oct/0310.html

Discussion:
Where to put this? Nowhere is there a strong section. If more than a few
paragraphs,
should just go to appendix of Reference document. Jeremy Carroll will
tweak. Needs to
be done in 2 weeks.

NEW ACTION: Jeremy to update his discussion of this issue, give quick
comments.
NEW ACTION: Guus to review, add to the Reference document as Appendix.
Guus: folks coming from RDF/S will move to OWL FULL. Make this clear.
Deb: a paragraph to go into overview, based on the longer discussion.
NEW ACTION: Deb to add a paragraph to Overview doc.
NEW ACTION: Jim, Pat to review

2. RESPONSE TO RDFCore LC (10 min)

ACTION Guus summary review of RDF Schema doc as proposal for the WG
->
The chairs will send a draft response to the list with a draft WG
response, including the detailed comments of Volz and ter Horst, the
XMLLiteral issue, and the (mixed) views on the social meaning issue.

Context:
Review by Volz of RDF Schema doc:
http://lists.w3.org/Archives/Public/www-webont-wg/2003Feb/0064.html
Review by ter Horst of RDF Semantics doc:
http://lists.w3.org/Archives/Public/www-webont-wg/2003Feb/0067.html
Review by Baget of RDF Concepts doc:
http://lists.w3.org/Archives/Public/www-webont-wg/2003Feb/0176.html

Discussion:
Pat: RDF core discussion group: subset of RDF/S and OWL-DL as extension
of this?
Pat: will keep WOWG up to date.
Jeremy: today is date for comments.
Jim: go through the comments, add additional comments.

1) Overall design issues
-Literals of RDF: domain of discourse needs to be redefined
-RDF XML literals as a datatype: drop this sentence from Jeremy?
Not a strong consensus
Peter: XML literal is a beast, makes OWL-lite very heavy
How much care needs to go into XML canonicalization and a particularly
well-specified
version of it to put into OWL-Lite.
Jim: contingent on RDF CORE doing the right thing.
Pat: different canonicalization doesn't break anything in the spec, just
underdetermined as
the XML spec is.
Jim: from pt of view of OWL Lite implementor: XML is just like an
annotation
Ian: but if a datatype, have to determine if two are equal
Jeremy: string compare.
Ian: different canonicalizers would cause different reasoners to come up
with different
responses.
Jim: rather, two different parsers could construct different
interpretations.
Jim: RDF will turn it into a string.
Peter: won't be a canonical string.
Jim and Dan: that is the point, turn into a canonical string
Pat: no, to take canonicalization under consideration

Jim: on design issue, drop paragraph

2) Parse datatype collections:
Jim: permit to be a list of literals.

3) Concepts doc:
Jim: concerns with social meaning issue; we have mixed views, no
consensus yet, some
serious reservations about RDF's description of social meaning.

Peter: two issues (RDF SCHEMA)
Jim: will reword the comments: consensus reached on notion in RDF
concepts doc,
because has an impact on OWL.

Pat: needs to be normalized.
Jim: the document should be clearer wrt the issue

4) RDF Schema document:
Raphael will send to RDF our comments.
Why is it called RDF Schema rather than Vocab. Spec. Language?
Domain and range constraints on super-properties with regard to
subproperties.
Raphael: in RDF Schema doc, if you add additional constraints to
subproperties, must be
compatible. No ambiguity in the formal semantics, but document is
unclear
WOWG agrees with Raphael's review. But Jeremy has some pts in section 4
that was
contrary to Raphael's.
Jim: we will exclude comments on section 4, not claiming consensus.

5) RDF Semantics doc:
Jim: we believe that the intended design of the semantics is such that
OWL will layer
appropriately, but need improvement to remove apparent inconsistencies.
Ter Horst has
reviewed. WOWG have asked Ter Horst and P-S to address semantics and
help RDF
semantics doc writers. Final doc fixes the inconsistencies.

Dan: how it is ok then? Or how is it broken?
Jim: the document has some problems, what wording should we use?
Ter Horst: observable as a test case, intentions are ok, but certain
errors in the document.
Jeremy: minor bug or semantic error, omission to be corrected.

ACTION Guus summary review of RDF Schema doc as proposal for the WG
PROPOSAL: this is our final version of the comments
PROPOSAL passes, resolved.
THIS ACTION ITEM CLOSED.

NEW ACTION: Jim to send comments to RDF CORE

3 ISSUETTES (45-55 min)

3.1 Actions wrt. resolutions

ACTION Jeremy to add 0542 case to test suite
CONTINUED

Action Carroll: XMLLiteral response to RDFCore
DONE:
http://lists.w3.org/Archives/Public/www-webont-wg/2003Feb/0194.html

ACTION Jeremy: to produce test case for XML literal
CONTINUED

ACTION Guus: explain AnnototationProperty in ref
CONTINUED

ACTION Guus: specify in owl.owl that label, seeAlso, isDefinedBy
DONE:
http://lists.w3.org/Archives/Public/www-webont-wg/2003Feb/0277.html

ACTION Mike S. will update guide to discuss annotations
CONTINUED.

3.2 OWL DL syntax

See:
http://lists.w3.org/Archives/Public/www-webont-wg/2003Jan/0489.html

Carroll's OWL Syntax doc:
http://lists.w3.org/Archives/Public/www-webont-wg/2003Feb/0146.html

Note from Jeremy:  the differences section highlights the crucial
differences everything else is either an orthogonal issuette, a bugfix
or a
trade-off.

Response by Peter:
http://lists.w3.org/Archives/Public/www-webont-wg/2003Feb/0161.html

See also: thread on complex restrictions:
http://lists.w3.org/Archives/Public/www-webont-wg/2003Feb/0266.html

DISCUSSION:
Jeremy: tried to simplify the concrete syntax of OWL-LITE and OWL-DL.
User wants to understand semantic constraints, what are the different
structures that have
patterns of other structures. Different syntactic patterns reduced to
URI issues. Blank
nodes are lists, data literals, restrictions, unnamed individuals, etc.
These concepts occur
as concepts in OWL. Blank nodes of specific types.  Mapping rules:
syntactic constructs
of restrictions. Doesn't see motivation in having in OWL-LITE. Simplify
variability of
other mapping rules.
Jim: what action?
Jeremy: Peter to be more open to Jeremy's suggestions to improve his
document.
Jim: do you have a list?
Peter: most of his suggested changes are not benefits.
Jim: how to actionize this?
Peter: 1) require abstract syntax move farther away from frame-base
syntax, otherwise
end up with a complicated mapping rule.
Jeremy: delete the mapping rule (for the complex restrictions)
Another fix: describe restrictions with multiple components in OWL-DL,
but only one in
OWL-LITE.
Dan: request to open an issue. Should this be on critical path. Chair to
have straw poll.

Jim: Straw Poll, because may take more time, so should this be an issue?

Support opening this as an issue?

6 in favor, 5 against, 6 abstain.
Chair to decide?
Jim: Guus, how do you feel?
Jim: ok, will open this as an issue.

NEW ACTION: Jim to open this as an issue.

NEW ACTION: Mike SMITH to put (message Jan 04089) in issues document.

Peter: 2) Jeremy claims to be making it easier for people, but will end
up with fewer
constructs, that is not a benefit.
Jeremy: type triple everything, but be more uniform
Dan: Sean, Jerome: implementation issues?
Dan: does it seem awkwardly irregular to you?
Sean: need to have a clean, abstract syntax, so if have to jump through
loops, is a
problem. Much prefer a clean syntax, more of a frame-form.
Jerome: problem that Jeremy has is more difficult, i.e., going from the
abstract syntax to
RDF triples.
Jos: the clean trick of Jeremy is important
Jim: as chair, how to reach consensus. Peter and Jeremy to reach
consensus first, and
others?
Ter Horst: RDF characterization of OWL-LITE and OWL-DL must be correct.
Missing
from the ASS document.
Jim: get a small group to discuss this and come back to the WOWG in a
week or so. Guus
to help out? Guus to take part of Jeremy's and part of Peter's and put
together.
Pat: clarification: provide RDF description for parser writers, also
expository issue to
describe OWL that makes it more compatible with RDF view. Latter: make a
document
that actually compares and contrasts the views.
Ian: anyone coming from RDF/S, will be actually in OWL-FULL world; this
seems
contrary to Jeremy's changes.
Jeremy: has a different view of use of OWL-LITE, from RDF to OWL-LITE,
so wants a
migration path.

NEW ACTION: Ian, Jeremy, Peter, Pat, Sean, Guus, ter Horst will work as
task force on
discussion and return with results in 1 week with progress. Probably
can't resolve this
next week.

Jeremy: probably the best we will get is understanding at the Plenary
Jim: have the group tell us what it has resolved, copy the group in your
discussion.

3.3 RDF compatibility

See:
http://lists.w3.org/Archives/Public/www-webont-wg/2003Feb/0083.html

Proposal to resolve by Carroll:
http://lists.w3.org/Archives/Public/www-webont-wg/2003Feb/0086.html

Discussion:
Patel-Schenider:
http://lists.w3.org/Archives/Public/www-webont-wg/2003Feb/0094.html

Relevant input: Carrol's OWL syntax note:
http://lists.w3.org/Archives/Public/www-webont-wg/2003Feb/0146.html

Note from Carroll wrt. the latter: In the differences section
   http://sealpc09.cnuce.cnr.it/jeremy/owl-syntax/2003-12-Feb/diffs.html

under RDF Compatibility/builtin names we see this issuette relates to
one
sentence of the mapping rules that excludes lots, and I instead propose
that
we allow everything unless it is prohibited by the table of urirefs

http://sealpc09.cnuce.cnr.it/jeremy/owl-syntax/2003-12-Feb/dl-syntax.html#builtin-

urirefs

3.4 ISSUE MOVED TO FIRST AGENDA ITEM

3.5 Imports and entailments

DISCUSSION:
Sean: the import mechanism doesn't support use of abstract syntax, or
makes it
complicated. Could have a syntactic construct spread over an ontology.
Jim: spent a long time on imports, don't hear a strong argument for
getting rid of it,
instead more specific. So no reason to change the status quo.
Sean: getting rid of it would be worse.
Jim: chair is resistant to reopening this issue.
Jim: bring this issue to a close.

Context: Carroll's message:
http://lists.w3.org/Archives/Public/www-webont-wg/2003Feb/0014.html

ACTION: Ian to get Sean to generate syntax checker test case for imports

(i.e., on failure assume doc is OWL-FULL)
DONE:
http://lists.w3.org/Archives/Public/www-webont-wg/2003Feb/0207.html

ACTION Jos: send negative experience on imports to list
DONE:
http://lists.w3.org/Archives/Public/www-webont-wg/2003Feb/0192.html

4. DOCUMENT STATUS (10-15 min.)

ACTION: Massimo publish reference (continued)
CONTINUED.

ACTION: Connolly get Tests published
DONE   http://www.w3.org/TR/owl-test/

ACTION: Connolly re: security
CONTINUED.

ACTION: Jeremy/HP review Reference
CONTINUED.

ACTION: Jim review Reference
DONE
http://lists.w3.org/Archives/Public/www-webont-wg/2003Feb/0264.html
OK TO PUBLISH

Jim: Media type? Dan to suggest reopening the issue?
Jim: MIME type. Mike to publish document with this incorporated?

NEW ACTION: Evan and Mike Dean to take MIME type and produce MIME type
doc.

Mike S: our web site doesn't have links to latest docs
Dan: yes, use "latest versions"

- Planning for LC candidate documents
- For all editors: include member list in ack section
http://lists.w3.org/Archives/Public/www-webont-wg/2003Feb/0263.html

Overview, Use Cases, Guide to add

6. AOB (0-5 min.)

Received on Sunday, 23 February 2003 11:32:58 UTC