Minutes for 05
September 2007 URW3
Meeting
Attendees:
- Anne Cregan
- Francis Fung
- Vipul Kashyap
- Mitch Kokar
- Kathryn Laskey
- Ken Laskey
- Thomas Lukasiewicz
- Eric Monk
- Matthias Nickles
- Mike Pool
- Peter Vojtas
Agenda:
- Approve minutes from previous two meetings. Note, the
minutes are linked from http://www.w3.org/2005/Incubator/urw3/#Minutes.
- Review status of action items from previous meeting
- Report from editing committee for final report. Draft
report at http://www.w3.org/2005/Incubator/urw3/wiki/DraftFinalReport.
- Review work plan and prepare to move to next phase of work
plan. Reference: http://www.w3.org/2005/Incubator/urw3/wiki/URW3WorkPlan.
- Discussion of cross-cutting issues from use cases.
- Other business
Scribe: Thomas Lukasiewicz
Discussion summary:
- Minutes from previous two meetings were *not* approved due
to insufficient quorum.
- Review status of action items from previous meeting
Action 1: Ken to check how to embed figures into Wiki page.
Status: Done: There are two scripts that allow embedding; they look
promising.
Action 2: Kathy to make a Wiki list of those attending ISWC workshop.
Status: Done. List at
http://www.w3.org/2005/Incubator/urw3/wiki/People_from_URW3_who_are_going_to_URSW_2007_in_Boursin.
Action 3:Kathy will send an email asking people who will attend URSW to
put their names on a Wiki.
Status: Done.
Action 4: Paulo, Trevor and Mitch will communicate among themselves to
do the organization of the editing team. For the next meeting, they
need to report the editing team strategy.
Status: Done; see also below.
Action 5: Ken will send an email asking who else volunteers to the
editing team while also announcing the current volunteers.
Status: Done.
Action 6: Giorgos will summarize RIF work relevant to us at a future
telecon.
Status: This is deferred till next meeting, because Giorgos couldn't
make today's meeting.
Action 7: Everyone add to Wiki discussion of sentence.
Status: The discussion has not been very active on Wiki. To avoid just
playing with words, we will focus on drafting text for final
report.
- Report from editing committee for final report. Draft
report at http://www.w3.org/2005/Incubator/urw3/wiki/DraftFinalReport.
- Editing committee is Paulo, Trevor, and Mitch.
- Mitch has sent us an Excel file with an editing plan.
- Editing team will coordinate a coherent writeup of use
cases. Each use case manager will write about the use case. We need to
develop a template for writing up use cases so they are all in a
consistent format.
- Mitch suggests that since we have an ontology, we should
review use cases for how well they cover the ontology.
- Ken says he tried to relate the Discovery use case back
to the then-current ontology. He doesn't remember how well we captured
this in the discussion. Ken said he would work with the Discovery use
case (http://www.w3.org/2005/Incubator/urw3/wiki/Discovery)
and work with the editing team to write it up in a good form. Ken and
Kathy and the editing team will confer offline in email about how to
write up the discovery use case.
- Peter has a principal problem: we describe the status of
web after our suggestions are used or about a web as of today? Ken says
part of our conclusions may be that a level of standardization will be
required. But the web will continue to be a diverse community. Kathy
says the use cases describe problems and a vision for their solution.
Kathy says this will help us to identify things that need to be
represented and therefore imply what needs to be standardized. Ken
reminds us that we won't write standards, but will lay out things we
think need to be standardized.
- Peter and others say that the process of RDFa annotation
is a main uncertainty problem. If you have uncertain information, you
can't put it on the web in RDFa without some attention to semantics and
some way of representing or resolving the uncertainty. Peter will help
editing committee on this point.
- Peter asks: If you have uncertain information, is this
problem related to machine understanding web data? Ken says need to
point to kinds of information you need and standard encoding that could
represent it. Example: SAWSDL does something trivial but critical --
gives you a way of pointing to semantic model that defines term. Link
could be just text, but could be updated in future to OWL ontology. The
idea: problem is understanding, first step is a standard that gives a
pointer to semantic models. This is a good template for us. We aren't
going to write a standard. We can identify things that are needed in
tags. And where it is needed. This gives us a way of thinking about
problems that can be solved. What do we need to capture and where -- we
can turn those conclusions over to somebody who can write the syntax.
Ken says the level of semantics we would specify is to say there needs
to be an annotation that points to a semantic model.
- Peter says that there is a difference between whether
something is annotated by the author or by somebody else afterwards.
Can be done manually or automatically. Matthias says that this might be
the issue of provenance annotation: annotate not just with a number but
with additional information such as author or other relevant
information. Ken says he is working on formal descriptions of services.
Ken says they are talking about 3rd party annotations. Peter might
create a service and annotate it with what he considers best practices
on using it. Then, Ken might read Peter's best practices, and then
certify it for his purposes. Then, Mitch might look at Ken's
certification, and it might be sufficient for Mitch because he trusts
Ken's certification process. Peter says that Mitch can observe his
agent has low precision - author's annotation is not sufficient. We
need 3rd party annotations. Mitch says this is a kind of indirect
reference or reification.
- Review work plan and prepare to move to next phase of work
plan. Reference: http://www.w3.org/2005/Incubator/urw3/wiki/URW3WorkPlan.
- Kathy says we are reasonably on target. Ken says we
still have a few weeks left where we can be discussing use cases. Also,
it took a while to get XG members signed up as Invited Experts and he
believes W3C would allow XG to slightly run over charter period if we
were finishing the XG report.
- Vipul asks what is a methodology? Vipul would call fuzzy
logic, probability, etc. a technology. He would say a methodology is a
way to match technologies to problems. A problem with semantic web
technology is how to make them useful to the masses.
- Ken says that if I'm making a doctor's appointment, I
should be able to say, "I am usually free on Wednesdays but usually
Tuesdays are booked up." And it should understand what that means.
- Anne suggests to annotate data with the appropriate
properties indicating which techniques might be applied:Adapting
existing systems (with their uncertainty technologies) for use on the
web. Maybe a wrapper would be needed to convert between different
uncertainty technologies. For example, the wrapper might say - random
variable, normal distribution - and then you know what kind of
statistics can be applied.
- Kathy gives a link to the presentation on combining
formalisms: http://www.galaxy.gmu.edu/QMDNS2007/WebPages/PRES/TCS1-1-Wright.pdf.
Kathy also mentions a paper by David Heckerman on the relationship
between certainty factors and probabilities from the 80's or early 90's.
- Discussion of cross-cutting issues from use cases.
- Anne mentions the POWDER Semantic Web Working Group: http://www.w3.org/2007/powder/.
- Anne mentions the Service Modeling Language (SML): http://www.w3.org/TR/2007/WD-sml-20070806/.
- Vipul mentions the Semantic Web for Health Care and Life
Sciences Interest Group (HCLSIG): http://www.w3.org/2001/sw/hcls/.
- Anne mentions the Semantic Web Coordination Group: http://www.w3.org/2001/sw/CG/.
- Other business.
- Next four sessions are September 19 and October 3, 17,
31. (We will have telecons every two weeks at this time.) Anne points
out that November 14 is ISWC.
Action items
- ACTION (1): Editing team to
go over IRC log and capture essential topics to include in our report.
- ACTION (2): Vipul
to relate health care use cases to HCLSIG activities.
- ACTION (3): Ken
and Kathy to join SW Coordination Group call on 14 September and will
check on relevance of POWDER, SML, possibly others to URW3 work.
- ACTION (4): Vipul
to relate HCLSIG use cases to uncertainty ontology.
- ACTION (5): Vipul
will give us a report at a future telecon.