RE: SUO: Re: SUO Ballot with 2 Questions
Dear Mike,
You wrote.
> Please spare me the false dilemma and histrionics. Why
> would the creation of a registry generate some magical
> collaboration of the SUMO and OpenCyc efforts? How will
> the existence of a registry increase the number of people
> working on the respective efforts?
MW: using a Registry approach is just a tool to enable
collaboration. It make it easier, so it is more likely to
happen. It is a good idea even if we just decided to develop
a single ontology.
>
> Secondly, even if it did make the contributors feel warm,
> fuzzy and collaborative, why is merger work between two
> candidates that have a great deal of conceptual overlap
> productive in efforts to generate a SUO?
MW: What I have seen suggests that (in the great scheme of
things) both OpenCyc and SUMO are immature. They lack hundreds
and thousands of manyears of work which is probably required
to achieve maturity. Therefore an environment in which different
ideas abotu an upper ontology can be placed and compared,
critiqued and improved will be helpful, as well as providing
a source of current best practice.
> We could vote to
> spend time registering and merging all the ontologies listed
> at the Bateman URL that you gave, I just don't see the point
> if our desire is to create a good usable standard in a
> reasonable amount of time.
MW: Well we can easily produce an ontology that about 20-30
people (maybe even a few more) think is a good and usable
standard in a reasonable time. We already have 2 (or 3 if
you count the EPISTLE stuff). However, building a larger
consensus is not about taking decisions, but about discussion
and reason around alternatives, and why one might work better
than another - given that the one thing we have is plenty of
alternatives. I personally think we would end up with more than
one ontology, but for good pragmatic reasons - there are applications
that would favour one style over another.
Matthew West
Principal Consultant
Shell Information Technology International Limited
Shell Centre, London SE1 7NA, United Kingdom
Tel: +44 20 7934 4490 Other Tel: +44 7796 336538
Email: matthew.west@shell.com
Internet: http://www.shell.com
> -----Original Message-----
> From: Mike Pool [mailto:mpool@iet.com]
> Sent: 04 June 2003 14:55
> To: John F. Sowa
> Cc: Patrick Cassidy; standard-upper-ontology@ieee.org
> Subject: Re: SUO: Re: SUO Ballot with 2 Questions
>
>
>
> At 09:27 AM 6/4/2003 -0400, John F. Sowa wrote:
> >Mike,
> >
> >You are getting hung up on an irrelevant detail. The
> >original version of motion #2 did not mention a registry:
> >
> >>But Patrick, this mechanism already exists within the
> >> proposed upper ontologies. I'm not sure how or why
> >> the registry provides a testbed that is not duplicated
> >> within an existing standard, but it's just not true that
> >> this registry proposal provides some capability for dealing
> >> with inconsistent theories that would not exist without
> >> the registry.
> >
> >The focus of the original proposal was to get three groups
> >that had not been talking to one another to work together.
> >Neither the SUMO nor the OpenCyc groups had any plans to
> >do anything like to the IFF work on relating ontologies,
> >detecting inconsistencies, and relating components to
> >one another.
> >
> >Adam Pease and I spent several weeks arguing back and
> >forth on the advantages of working with the IFF and
> >OpenCyc groups. After all that talk, Adam finally
> >admitted that maybe such a collaboration might not
> >"block" his own plans for further SUMO development.
> >
> >The idea of the "registry" was recommended by Frank Farance
> >and Matthew West, both of whom have had a great deal of
> >experience with international collaborations on standards.
> >They recognize that any major project needs some systematic
> >way of recording who did what, when, where, why, and with
> >what starting resources.
> >
> >That is the purpose of the registry: best practices
> >for software development and organization applied to
> >the task of ontology development and maintenance.
> >
> >If motion #2 fails, we will go back to status quo ante.
> >If you recall, the SUO mailing list had been very quiet
> >with very few emails per month. Then 5 things happened:
> >
> > 1. John De Oliveira submitted a motion to make OpenCyc
> > a candidate for a standard.
> >
> > 2. Adam Pease resubmitted his motion from two years ago
> > to make SUMO a candidate for a standard.
> >
> > 3. Then I suggested that we bring all three efforts -- IFF,
> > OpenCyc, and SUMO -- into a collaboration instead of
> > a competition.
> >
> > 4. John D. and the other Cyc members agreed that was
> > preferable to having independent efforts.
> >
> > 5. But I had to spend long email debates with Adam before
> > he was willing to join the coalition -- and then only
> > on the condition that he could have his own separate
> > motion.
> >
> >Now motion #2 proposes that we actually begin the work of
> >a community effort rather than several isolated efforts.
> >
> >To put it bluntly, a vote against motion #2 is, in effect,
> >a vote to dissolve the SUO Working Group. If it fails,
> >SUMO will go back to being a Teknowledge effort with a few
> >volunteers making comments. OpenCyc goes back to a CYc
> >effort with a few volunteers making comments. And IFF goes
> >back to a theoretical effort with no content to work with.
> >
> >If you can think of any better way to get the SUO group to
> >work together, then please suggest it. Otherwise, you can
> >vote for motion #2 or vote to dissolve the SUO.
>
> John,
>
> Please spare me the false dilemma and histrionics. Why
> would the creation of a registry generate some magical
> collaboration of the SUMO and OpenCyc efforts? How will
> the existence of a registry increase the number of people
> working on the respective efforts?
>
> Secondly, even if it did make the contributors feel warm,
> fuzzy and collaborative, why is merger work between two
> candidates that have a great deal of conceptual overlap
> productive in efforts to generate a SUO? We could vote to
> spend time registering and merging all the ontologies listed
> at the Bateman URL that you gave, I just don't see the point
> if our desire is to create a good usable standard in a
> reasonable amount of time.
>
> Mike
>
>
> >John
> >
>
>