SUO: Re: IFF Example?
o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o
Matthew et al.,
Let me suggest a working sttategy just to get our feet on some ground.
Pick an object in an object domain, say, a "triangle in geometry" (TIG).
All of the things that Matthew would say about a triangle in geometry make up
"Matthew's theory of a triangle in geometry", Theory (TIG, Matthew), or we
could say that all that all of the properties that Matthew would predicate
about a TIG make up "Matthew's comprehension of a triangle in geometry",
say, Comp (TIG, Matthew).
All of the things that Matthew would recognize as a triangle in geometry make up
"Matthew's denotation or extension for the concept or term 'triangle in geometry'",
say, Deno (TIG, Matthew).
o Comp (TIG, Matthew) = {properties)
'
' o Theory (TIG, Matthew) = {sentences}
Geometric Figure o ' "
/|\ ' "
/ | \ ' "
/ | \' "
Point o o o Triangle
Edge ' "
' "
' "
' "
' o Concept or Term "Triangle In Geometry"
'
o Deno (TIG, Matthew) = {objects}
The trouble is, and the trouble has been, that agents disagree,
not just in the particular labels that they happen to use, say,
the way that English, French, and German geometers might use
different words to say the same things about triangles, but
disagree in some radically fundamental and structural way.
We can use the concept of "functors as facets or views"
to think about these sorts of situations. Say that we
all share the hypothesis that there is some objective
reality to be gotten at behind a slected set of words.
Call this object or objective domain "X". Then we can
think of each agent's view of X, individual or communal,
personal or embodied in a formalized or computable form,
whatever you please, to be a projection mapping from X
to the agent's own conceptual or terminological space,
say, v_i : X -> E_i for agent i.
X
o
/|\
/ | \
v_i / ... \ v_j
/ | \
/ | \
v v v
E_i ... E_j
Information-wise, all of your non-divine agents are bound
to have partial views of X, so each v_i : X -> E_i loses
some part of the information that would be necessary
to describe X X-haustively.
We can now express our problem, just between us non-divine agents,
as one of relating the various partial views of X into something
more complete, which will involve ironing out many inessential
wrinkles around various points of singular inconsistency.
Jon Awbrey
o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o
West, Matthew R SITI-ITPSIE wrote:
>
> Dear Robert,
>
> Count me in as prepared to put some sample EPISTLE material in the
> pot and support the analysis to map to some other ontology.
>
> 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: Robert E. Kent [mailto:rekent@ontologos.org]
> > Sent: 06 May 2003 17:50
> > To: Patrick Cassidy
> > Cc: SUO; John F. Sowa; Leo Obrst; Jim Farrugia; Marco Schorlemmer
> > Subject: SUO: Re: IFF example?
> >
> >
> >
> > Patrick and others,
> >
> > ----- Original Message -----
> > From: "Patrick Cassidy" <pcassidy@bellatlantic.net>
> > Cc: <standard-upper-ontology@ieee.org>
> > Sent: Friday, May 02, 2003 8:41 AM
> > Subject: SUO: IFF example?
> >
> > [snip]
> >
> > > The
> > > task of organizing logically inconsistent ontologies --
> > > or John's lattice of theories -- is also important, but
> > > I would like to first focus on the task of finding
> > > translations ("mappings"?) between logically consistent
> > > ontologies (or portions of them).
> > > Is it possible for someone working with IFF to
> > > prepare a very simple ("as simple as possible, but
> > > no simpler") example of how IFF could be used to
> > > translate between a small fragment of two ontologies,
> > > say between OpenCyc and SUMO? If an IFF-er is
> > > willing to try, I would be happy to make suggestions
> > > as to what areas to try to translate.
> >
> > Marco Schorlemmer www.dai.ed.ac.uk/~marco and I would be
> > happy to work with
> > you and other interested people on the issue of how the IFF
> > can serve as a
> > framework for relating the modules of SUMO, OpenCyc and other
> > object-level
> > ontologies, and for organizing their modules and microtheories into a
> > lattice of theories. Neither of us are experts on either SUMO
> > or OpenCyc,
> > but we are certainly interested in the issue of ontology
> > mapping. We are
> > particularly interested in using the application of
> > Information Flow-based
> > ontology mapping methodology to convince the SUO WG of the potentially
> > practical value of the approach taken in the IFF.
> >
> > Robert E. Kent
> > rekent@ontologos.org
o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o~~~~~~~~~o