Thanks, Mike. =ppy
-- (01)
--- Begin Message ---
At the end of this, is my suggestion for a way forward. (01)
Mike
------ (02)
1 UO Summit Telecon - 03/2/2006
Steven Ray: what are goals? (03)
1. That there is a way to move forward through the mess of UOs,
there will be a path to link in with other commercial systems etc. This
is the primary signal to be sent out. Will help commercial organizations
as well as standards organizations.
a. We first have to convince ourselves that there is a way to
transfer information between the different UOs.
b. Identify most promising avenues for
moving forward. (04)
Public Straw Man announcement: (05)
BillA: there is a lot of enthusiasm here. There needs to be empirical
data on what actual uses of UOs in commercial situations. Talk of
merging the UOs is premature. (06)
PatC: Agree about need for data, but large companies may not commit till
there is more acceptance. (07)
BillA: to be an alternative to the semantic web, need backup. (08)
MattW: in oil industry, we have large deployments of our UO, simple use
of common language to organize design data, no inference. (09)
SteveR: Lenat would argue that he has industrial strength applications
built on top of Cyc. RalphH of TopQ speaks of some commercial
applications. (010)
MillsD: that can be topic of discussion. (011)
SteveR: action item before the meeting: actual industrial studies. Need
a compelling case for public to adopt. (012)
PatC: can you tell us who your customres are at OntologyWOrks? (013)
BillA: our customers don't want their use cases made public. We have one
direct study that using our UO added value. (014)
SteveR: Can custodians commit to identifying the similarities and
differences and links between their different UOs. Make explicit the
different distinctions and assumptions underlyign the different UOs.
This is critical for users to decide.
Can only xfer those things that are compatible. Map the landscape of
UOs, showing what is included or exclude by each UO. (015)
MattW: must also map equivalent concepts.
Leo: gets into intension/extension issue. (016)
DagS: there might be specifications of time that are only in one
ontology, not the others. Label things with their origin, pick and
choose from different UOs. (017)
SteveR: lets commit to making a map of the different UOs. Identify scope
so can easily be compared. (018)
Leo?- easier at lowere or middle levels.
SteveR: that is Lenat's position. (019)
SteveR: I want to move towards being able to communicate to people what
the value of ontologies and upper ontologies is. (020)
PatC: We need to agree that we are looking for relations between the
different upper ontologies. Then present to potential funders. (021)
SteveR: subsets may be funded projects.
DagS: I agree, (022)
Nicola: put bullet on agenda for funding.
Nicola: Add to steveR's comment to document the design rationale for the
various UOs. We can explore this landscape in various ways. One, compaer
ontologies as a whole, as logical theories, etc. Also need to go into
details of comparing single concepts, relations etc. (023)
SteveR: next itme on agenda. Engaging custodians. Need cooperation by
owners of the UOs. How to get all of them engaged? (024)
Nicola: are there any custondians against our vision? (025)
SteveR: there is skepticism, it may not succeed. Some not convinced of
value, most of the value is in the middle and lower levels. (026)
Leo: we know Doug and Adam have not yet participated. Lets ask them to
participate or designate someone to do so. (027)
Steve: Maybe send along some of our proposed goal/vision statements to
get reaction. (028)
PatC?: Adam is away, he is interested. (029)
DagS: comment on imtermediate ontologies. One idea is to use formal
thinking of upper ontologies to make middle level ontologies more
correct. (030)
Many people have middle ontologies.. thre is possibility that getting
more understanding of UOs and their relationships, could help people (031)
PatC: what doug is saying when he says UOs don't matter, he is saying
the specific UO you use matters much less than using A UO. (032)
SteveR: Lenat may say not even matter if we use same UO, as long as we
can do business together. (033)
NIcola: agree that middle level stuff really matters. UO matters when we
need to talk to each other or need to do business together. (034)
MIkeU: Another benefit of a UO is reuse, not have to redo tings over and
over again. E.g. a desk is a physical object, now you get all the axioms
of physical object, not have to redo. (035)
ACTION: collect benefits of UOs, contact each custodian as a starting
point. (036)
Nicola: ECOR... originally a Barry Smith creation, moving to Buffalo,
Nicola being asked to take over. Actually has no official status, like
NCOR to some extent. (037)
PatC: Actualy, NCOR is getting a bit of funding.
Nicola: I did not realize that, ECOR is not getting funding. (038)
Funding Possibilities: JohnWalker, Nancy Lawler,... (039)
PatC: people say it is not my department, saving $100M in
interoperability costs... (040)
StevR: NIST's job is to bring benefit to Industry. (041)
Leo: can we identify the contentious issues to address on the 14th? (042)
MIkeU: overall picture / way forward:
* critical first step is to map the landscape of current UOs
* next: identify feasibility and desirability of
merging/mapping/relating the different UOs
* if s feasible, then identify various ways forward, pick one or
more, and go.
* If not feasible, then carefully document why and go acknowledge
that this is not gogin to happen for a while, if ever. Then decide what
can be rescued from what we have, and recommend a way forward for UOs. (043)
MikeG: this is fantastic, make sure to have pointers to the axioms in
the individual UOs. (044)
Is the map the landscape task going to be done/started before the
meeting? (045)
MikeU: great idea to get started. Each customdian summarize their own UO
in basic terms like:
* what is it for?
* What are its basic assumptions?
* In your own words, how to you describe the key elements of your
UO? (046)
--- End Message ---
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/uos-convene/
To Post: mailto:uos-convene@xxxxxxxxxxxxxxxx
Community Portal: http://ontolog.cim3.net/
Shared Files: http://ontolog.cim3.net/file/work/UpperOntologySummit/uos-convene/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?UpperOntologySummit (01)
|