| Thread Links | Date Links | ||||
|---|---|---|---|---|---|
| Thread Prev | Thread Next | Thread Index | Date Prev | Date Next | Date Index |
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-----At 10:33 AM 6/6/2003 +0100, West, Matthew R SITI-ITPSIE wrote:
From: Mike Pool [mailto:mpool@iet.com]
Sent: 06 June 2003 14:02
To: West, Matthew R SITI-ITPSIE; Eric Peterson; John F. Sowa; apease@ks.teknowledge.com; clegg@cyc.com; John DeOliveira
Cc: Patrick Cassidy; standard-upper-ontology@ieee.org
Subject: RE: SUO: Re: SUO Ballot with 2 Questions
Dear Eric,
Perhaps I am just a pragmatist, and will settle for what can be
done rather than go for what I know cannot be done.
What cannot be done (today or in the next 5 years) is to create
a single canonical ontology that will gain universal, or even
very widespread acceptance. The reason for this is simple, some
18 months ago there was considerable discussion here about 3D
and 4D approaches, which concluded that they were contradictory,
and so could not be merged in the way that different ontologies
that had the same basic assumptions could be. If you look at
Nicola Guarino's paper:
http://wonderweb.semanticweb.org/deliverables/documents/D15.pdf
You will see some other choices that different ontologies make
also identified, where there can be good reasons for particular
choices depending on your objectives. There is therefore no
prospect of widespread agreement on a single ontology, there
are just too many different camps with significant support.
The question in my mind then is how do we keep people working
together towards one goal, rather than suffer fragmentation.
What I think can be achieved is a modest number (say 2-4) of
ontologies with different foundational principles, that are
mapped to each other and that incorporate merged elements from
a number of ontologies of their persuasion. By being inter-
related, they can even be thought of as a single ontology. An
assertion in terms of one foundation will have implications in
the others.
Matthew,
I wholeheartedly agree that a SUO must be capable of allowing for different representational approaches and the co-existence of ontologies with very different foundational principles is justifiable, at least in the short term until they can be merged into a single ontology with distinct modules but sharing terms where appropriate..MW: I too agree this is part of where we should be going.
But note what was being proposed. It was being proposed that the SUMO and OpenCyc be put in this registry. This isn't a case of two ontologies representing radically different approaches to, for some representational issue, e.g., the 3D vs. 4D problem, neither of which we can do without. This is a case of two ontologies with very similar foundational principles.MW: I think the issue is more that they don't have clear foundational principles. I see putting them in a registry as the first step in a bottom up approach to sifting and sorting them to find the implicit foundational principles, or establish some and modify and merge bits that don't conform.What you're saying now seems reasonable but this is not how the registry was being proposed. It was, as you acknowledged, more of a political move. Here's what you said on May 23 in a note to Pierre Grenon:
- I fear you are being far to logical.
- In the end we are just recognising some initiatives as part of the
- work of this group. This is largely a political exercise rather
- than a technical one. Strictly, you could wait with any of these
- initiatives until you had a complete standard, and then put
- it forward in a motion to be accepted as a standard.
MW: I was refering above to the matter of having some motions and voting, rather than the content of the motions.Now you're reintroducing the registry as essential to completeness and restricted only to a very small number of ontologies with very different foundational principlesMW: My words would be "very useful" rather than "essential". I am sure you coudl do it other ways, it would just take longer, be more expensive, lower quality etc. Nor am I usggesting restriction to a small number of ontologies. That is just my guess on how it will turn out. We should not be trying to constrain the outcome, but let it emerge.
I remain confused about the function and objectives of a registry.MW: I've said it before, but it perhaps bears repeating. It is primarily a tool to help us manage what we are doing. It does not force us into any of the alternatives of one vs many ontologies, but I assume we wish to minimise divergence, and it should help us to do that.
best regards,
MikeJohn's motion sets the scene for this to happen.
I guess we will find out how the land lies when we get the result
of the vote.
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: Eric Peterson [mailto:epeterson@CCAAVA.com]
> Sent: 05 June 2003 15:15
> To: West, Matthew R SITI-ITPSIE; John F. Sowa; Mike Pool;
> apease@ks.teknowledge.com; clegg@cyc.com; John DeOliveira
> Cc: Patrick Cassidy; standard-upper-ontology@ieee.org
> Subject: RE: SUO: Re: SUO Ballot with 2 Questions
>
>
> Hi Matthew;
>
> Please see below:
>
> > -----Original Message-----
> > From: West, Matthew R SITI-ITPSIE [mailto:matthew.west@shell.com]
> > Sent: Thursday, June 05, 2003 5:26 AM
> > To: Eric Peterson; John F. Sowa; Mike Pool;
> apease@ks.teknowledge.com;
> > clegg@cyc.com; John DeOliveira
> > Cc: Patrick Cassidy; standard-upper-ontology@ieee.org
> > Subject: RE: SUO: Re: SUO Ballot with 2 Questions
> >
> > Dear Eric,
> >
> > Why do you think it is contentious to manage the content of the
> > lattice of theories? (which is all that using a registry means
> > in practice?)
>
> [ELP] That's really not what I said.
>
> We are contending on the issue of registries. It is, therefore, by
> definition, a contentious issue. That's all I was saying. You would
> think we were speaking two different languages that diverged
> a couple of
> hundred years ago ;^)
>
> But as to your concern, if you or I had proposed resolution #2, this
> discussion would not still be happening. But John has earned
> the right,
> by virtue of his credentials and reputation, to have his
> motions stay on
> the table longer than they otherwise might. I don't have a
> problem with
> that.
>
> But do we want to turn into a group that refuses to do something as
> simple as state as a goal that as a SUO group we are going to merge
> worthy ontolgies into an upper ontology. We are not the SUOs
> group (the
> pluralization of SUO). The direction of this group is
> obviously getting
> warped off of its original intent.
>
> Motion #2 elevates the notion of the organization of competing efforts
> while refusing to commit to work on merging them.
>
> That is the problem I have with it.
>
> If the registry is really just a tool to help us merge lets say that.
> If we don't believe in merging, then I offer a drastic, but needed
> solution below.
>
> I actually, prefer that we choose one ontology as a starting point and
> fold in worthy additional content as we are able. I question
> whether we
> have the unity or enough spare time as a group do a wholesale
> merge very
> successfully.
>
> But I offer merging as a compromise because it just strikes
> me as absurd
> to call a smorgasbord a standard meal - to borrow a metaphor.
>
> But if Constantine, at Nicea, could leverage a compromise where three
> became one, how much easier to say that two are one ;^) It
> will become
> a beautiful mystery of SUO faith how the SUMO and the OpenCyc are
> actually one standard upper ontology ;^)
>
> Sorry out there, but the parallel was just too striking and I couldn't
> help myself.
>
>
> >
> > Please also note that SUMO and OpenCyc are not the only potential
> > contributions.
>
> [ELP] I was addressing the current motion. Robert's rules certainly
> allow for subsequent motions. I expect that we could gain
> much from the
> others.
>
> >
> > If a single ontology is the "right" answer, putting the constituent
> > ontologies into a lattice and asking questions about the differences
> > is probably the best way to bring that about.
>
> [ELP] If that is what we are doing, we need to say that in the motion.
> But maybe we need to split into an SUO and an SUOs.
> Socially, it would
> be a lot of fun for us all to get together and have a barbeque or
> something. But technically, I'm afraid our goals may be too
> dissimilar
> to allow for anything but spirited debate and obnoxious religious
> metaphors.
>
> I officially move that we "choose ye this day" (or some day soon)
> between one SUO and an SUO of SUOs.
>
> If someone wants to formalize that last sentence, I would
> take that as a
> "second" of the motion.
>
> <snip>
>
> FWIW,
>
> -Eric
>