ontolog-forum
[Top] [All Lists]

Re: [ontolog-forum] Foundation ontology, CYC, and Mapping

To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: Jack Park <jackpark@xxxxxxxxx>
Date: Wed, 17 Feb 2010 15:17:07 -0800
Message-id: <5179aafa1002171517h722a7e2ardf170db6290fc2bb@xxxxxxxxxxxxxx>
No objection raised for Peter to host the project; the remaining
question, it seems to me, would be whether Peter has the facilities to
host a public SVN. I think he does in relation to some of his wiki
work.    (01)

Jack    (02)

On Wed, Feb 17, 2010 at 2:49 PM, John F. Sowa <sowa@xxxxxxxxxxx> wrote:
> Jack,
>
> Thanks for the note of support.
>
> I believe that the only way to get this group to do anything is to
> start with something fairly good and build from there.  OpenCyc is
> the most likely candidate, but the FO should give equal support to
> other reasonable ontologies.  Therefore, it should be hosted on
> neutral grounds, rather than on a web site that is devoted to
> a specific ontology, such as OpenCyc.
>
> Re Google:  That is a possibility, but Peter Yim has been doing
> a good job in hosting ontolog forum and the related web sites.
> Therefore, he should have the right of first refusal to host
> whatever this group develops.
>
> John
>
> -------- Original Message --------
> Subject: Re: [ontolog-forum] Foundation ontology, CYC, and Mapping
> Date: Wed, 17 Feb 2010 12:17:53 -0800
> From: Jack Park <jackpark@xxxxxxxxx>
> Reply-To: [ontolog-forum]  <ontolog-forum@xxxxxxxxxxxxxxxx>
> To: [ontolog-forum] <ontolog-forum@xxxxxxxxxxxxxxxx>
>
> I would toss a +1 at this concept of starting somewhere and evolving
> from there. OpenCyc makes sense. What else makes sense might be a
> website for such a project,  for instance, http://code.google.com/ ,
> which provides Subversion (and other) version control options, a wiki,
> download area, etc. Perhaps there are other options.
>
> Jack
>
> On Wed, Feb 17, 2010 at 8:41 AM, John F. Sowa <sowa@xxxxxxxxxxx> wrote:
>  > Dear Matthew, Doug, and Pat C,
>  >
>  > Before getting to the details of your notes, I'd like to clarify the
>  > term FO, which Pat introduced for a Foundation Ontology.  I have no
>  > objection to using the term FO, but I would apply it to the totality
>  > of the theories (or microtheories) in an OOR.  The Cyc example of an
>  > upper ontology with many specialized microtheories would be an example.
>  >
>  > However, I would broaden the idea to support multiple theories at the
>  > upper levels, which might be incompatible.  For example, Matthew,
>  > Chris P, and Pat H have strongly supported a 4D ontology for the upper
>  > levels, but many other people prefer to use a 3D upper level.  For
>  > many of the lower level microtheories, the differences between a 3D
>  > vs 4D foundation are irrelevant.
>  >
>  > Pat C wants to use primitives based on Longman's dictionaries, but
>  > I objected that those terms are too vague, and an enormous amount
>  > of work would be needed to make them precise.  Therefore, I propose
>  > the following approach:
>  >
>  >  1. OpenCyc is freely available as open source, but Pat objected
>  >     that it has very few axioms.  However, Longman's terms have zero
>  >     axioms, and they have never been tested for their usefulness in
>  >     any ontology, certainly not in anything as large as Cyc.
>  >
>  >  2. Therefore, I recommend that we adopt the OpenCyc terms together
>  >     with the axioms available in OpenCyc as a "starter set" for
>  >     developing the Foundation Ontology.
>  >
>  >  3. But we would also welcome and encourage anybody with different
>  >     preferences to contribute more terms and theories to the FO.
>  >
>  >  4. When convenient, those new theories should adopt the same
>  >     spelling as the terms already in the starter set, but any
>  >     new terms (taken from Longman's or any other source anyone
>  >     might prefer) could also be used in those additional theories.
>  >
>  >  5. The complete hierarchy of theories for the FO would be
>  >     organized along the lines we've been discussing in these notes.
>  >
>  >  6. Pat's goal of finding a much smaller set of defining terms
>  >     (called primitives) could be done in parallel with the
>  >     development of the FO instead of delaying its development.
>  >
>  >  7. If and when a good set of primitives has been found useful,
>  >     a methodology based on them could be promoted for simplifying
>  >     further ontology development.  However, older theories that
>  >     use the older terms would remain available as long as anyone
>  >     needs them.
>  >
>  > I'm suggesting this as a compromise that would begin with a large
>  > tested set of terms already organized in a generalization/
>  > specialization hierarchy.  It would also accommodate new terms
>  > that could be added as needed from any source.  But if anyone
>  > wants to do any immediate implementation today, the OpenCyc
>  > terms can be used.  Any work done with them would be guaranteed
>  > to be supported by the full FO.
>  >
>  > MW> I think [the FO] is more than just a vocabulary, but I agree
>  >  > that great care would need to be taken not to introduce into what
>  >  > I am calling abstract theories axioms that did not contradict say
>  >  > the 3D and 4D axioms that would be introduced when they were
>  >  > combined with those theories.
>  >
>  > I agree.  In fact, one advantage of OpenCyc is that it doesn't have
>  > all the axioms (AKA old baggage) of full Cyc.  That implies that it
>  > is more general than Cyc and less likely to have unpleasant
>  > "surprises" (AKA inconsistencies).
>  >
>  > MW> However, having such [abstract theories] would greatly simplify
>  >  > mapping between say 3D and 4D ontologies. These would I think need
>  >  > to be carefully designed rather than just picking something up,
>  >  > since it is casually almost certainly made some upper ontology
>  >  > assumptions.
>  >
>  > I agree.  Many, if not most of those theories could be based on
>  > terms that are already in OpenCyc.  In fact, if you just adopt
>  > the terms by themselves without any axioms, there is no possibility
>  > of an inconsistency.  Then any axiom that is added could be tested
>  > for consistency with both the 3D and the 4D ontologies (along the
>  > lines that I suggested in my previous note).
>  >
>  > DF> An ontology with a single theory would complicate the idea of
>  >  > a foundational ontology as an interlingua to which all external
>  >  > ontologies can be mapped.
>  >
>  > Yes.  That is why I would prefer to use the name Foundation Ontology
>  > for the full hierarchy of all the theories in the OOR.
>  >
>  > DF> Competing external theories could be incorporated by defining new
>  >  > concepts/relations for the relata of the competing theories (which
>  >  > map directly to the terms of the source ontology) and then adding
>  >  > rules relating them with the ostensible "sameAs" terms already in
>  >  > the foundational ontology.
>  >
>  > PC> Yes, but I would express the process as "logically specifying
>  >  > the terms of each extension ontology using only the terms in the
>  >  > FO" rather than "map directly" since that suggests that the
>  >  > entities in the extension ontology are already on the FO, which
>  >  > in general they will not be.
>  >
>  > I would expect the FO to be completely open-ended so that the
>  > number of useful terms would increase indefinitely.  There could
>  > also be a small subset of recommended defining terms.  (I wouldn't
>  > even object to calling them "primitives".)
>  >
>  > But we do have to recognize that many "primitives" may be very
>  > underspecified.  For example, the term PointInTime should be neutral
>  > with respect to a 3D or 4D ontology. For lower-level microtheories,
>  > you might have a general theory called Hiking, which would use
>  > PointInTime without any dependencies on 4D or 3D.  But if there are
>  > dependencies, one could have common specializations of Hiking with
>  > either of the two upper-level theories to generate the subtheories
>  > Hiking4D and Hiking3D.
>  >
>  > JFS> But if you think of [the FO] as a collection of theories organized
>  >  >> in a hierarchy, no single theory ever changes.  Instead, each
>  >  >> innovation adds another theory to the hierarchy, which may be a
>  >  >> generalization, a specialization, a sibling, or a cousin of some
>  >  >> other theory.  You  can also compare and combine theories.
>  >
>  > DF> Agreed.  This is part of the reasoning behind Cyc method of
>  >  > "microtheories" or contexts....
>  >
>  > PC> And that would be true for the FO and its extensions also. But there
>  >  > is an additional aspect to the FO as I have proposed it.  The CYC
>  >  > BaseKB is part of every other more specialized microtheory, but it
>  >  > was not designed as, nor used as, an inventory of basic elements that
>  >  > is sufficient to specify (as combinations) the intended meanings of
>  >  > the symbols that are in all of the other linked microtheories.  CYC
>  >  > didn't try that tactic, which could have been very informative.  But
>  >  > a proper test would in any case require that a good number of separate
>  >  > groups with different applications and viewpoints try to use the same
>  >  > FO to describe their different domain ontologies.
>  >
>  > This comment mixes two different goals:
>  >
>  >  1. A large useful ontology that people can begin using ASAP.
>  >
>  >  2. A project to determine whether a small subset of terms (called
>  >     primitives) is sufficient to define everything else and would
>  >     thereby promote interoperability.
>  >
>  > Different people may have different priorities.  My personal preference
>  > is to emphasize #1.  I have no objection to anyone who prefers #2, but
>  > I would not want to tell people who have a day job that they have to
>  > wait until #2 is finished.  My recommendation is to start with #1 and
>  > let anyone who wants #2 to *extract* some subset from #1 in order to
>  > test that hypothesis.  If that hypothesis seems to be justified, then
>  > the results could be developed into a methodology for using, adapting,
>  > and streamlining the much larger resources of #1.
>  >
>  > But I would never suggest that people who need an ontology today should
>  > wait until project #2 is completed.  The above proposal lets people
>  > start new projects using OpenCyc and rest assured that the FO would
>  > continue to support them in the future.
>  >
>  > JFS> You can think of context as some additional statements S that are
>  >  >> added to a theory T to specialize it for some particular application.
>  >
>  > DF> If the context referred to is what Cyc calls a DataMicrotheory,
>  >  > then the statements added are qualitatively different from those
>  >  > in the basic theory.
>  >  >
>  >  > A context might close T's open world assumption, such that T2 has
>  >  > a closed world assumption.  T and T2, in such a case, would be
>  >  > different types of theory.
>  >
>  > That's a good point.  It illustrates an important advantage of
>  > starting with OpenCyc (or some subset of it).  We can take advantage
>  > of the 26 years of experience in developing Cyc.  We don't have to
>  > adopt every one of their decisions, but when we diverge, we should
>  > have a good reason for doing so.
>  >
>  > DF> An FO would need to have a reasonably restrictive generalization
>  >  > of classes included in any microtheory that is to be mapped to it.
>  >  > Defining "reasonably restrictive" could be hard, but it seems to me
>  >  > that SUMO (with extensions) and Cyc both would currently meet this
>  >  > requirement.  Including concepts from UMLS and GoodRelations would
>  >  > lower the rough lower edge of the directed acyclic graph of classes
>  >  > in several key areas.
>  >
>  > I'm all in favor of building on good work that has been done in other
>  > systems.  We should make sure that the licensing terms are compatible
>  > and get explicit permission to use whatever is incorporated in the FO.
>  >
>  > DF> One question is on what basis should individuals should be included
>  >  > in an FO.  Certainly units of measure should be.  Currencies &
>  >  > countries surely.  Cities and every instance in the GeoNames base?
>  >  > How should the selection of people to add be made?  Organizations?
>  >  > Conceptual works (books, movies, songs, albums, paintings,
>  >  > constitutions, poems, ...)?  Sports and games? Events (disasters,
>  >  > wars, elections, mergers, ...)?  Etc.
>  >
>  > Those are excellent questions.  I would prefer to err on the side of
>  > being more inclusive.  My suggestion would be to keep the axioms
>  > relatively free of individual names, but to have an associated
>  > database that would store as much as anyone might find useful.
>  >
>  > However, there will undoubtedly be many very useful specialized
>  > theories, such as US IRS Tax Code for 2010.  The most qualified
>  > people to develop such a theory would be the IRS.  But the FO
>  > could maintain pointers to such theories stored and maintained
>  > in compatible formats by other organizations.
>  >
>  > DF> The breadth of coverage of the proposed FO needs to clarified.
>  >  > Is it to be an ever-expanding set of all terms defined in any
>  >  > ontology?   Should it include all individuals ever defined on
>  >  > the Semantic Web?
>  >
>  > Those are important policy decisions.  Since the SW is expanding
>  > very rapidly, we can't hope to incorporate it into the FO.  But we
>  > must have interfaces to it that would allow any application to
>  > access it as needed.  In fact, if we do a good job on the FO,
>  > the SemWebbers might take notice and adapt their technology to
>  > facilitate sharing in both directions.
>  >
>  > DF> Or could there be a basic, relatively fixed, FO to which an
>  >  > expanding number of contextually restricted, but still centralized,
>  >  > ontologies are related?  I could see such for brand-name products,
>  >  > GeoNames, UMLS, IMDB, GeneBase, etc.
>  >
>  > PC> The FO itself should try to include all primitives that are used
>  >  > by more than a small set of specific domain ontologies, and only
>  >  > those non-primitive elements that are needed for ease of use and are
>  >  > non-controversial.  Primitives required by domain ontologies should
>  >  > also be maintained, but as part of a domain extension.
>  >
>  > The goal of getting a useful FO ASAP implies that we should start
>  > with a much larger set of terms than Pat has in mind.  But that may
>  > be an advantage.  The goal of extracting a smaller number of defining
>  > terms can be guided by usage patterns.  Those terms that are most
>  > widely used would be prime candidates.  The other terms could be
>  > redefined in terms of them.
>  >
>  > I would expect an FO committee to be similar (in some ways) to the
>  > W3C.  It would define formats, guidelines, policies, etc., and
>  > encourage other groups to adopt them and make their work compatible.
>  >
>  > John
>
>
> _________________________________________________________________
> Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
> Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
> Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
> Shared Files: http://ontolog.cim3.net/file/
> Community Wiki: http://ontolog.cim3.net/wiki/
> To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
> To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx
>
>    (03)

_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (04)

<Prev in Thread] Current Thread [Next in Thread>