Re: SUO: Re: (ELP's summary of MRW's standards experience) Was: A NEW FUNDAMENTALLY DIFFERENT FORMAL MOTION...

["John F. Sowa" <sowa@bestweb.net>]

Erik,

To be clear, I would be very happy if we could develop
an ideal SUO which could support everybody's pet theories
as special cases.  Since I don't believe that is likely to
happen in the next couple of years, I think that the best
we can do, in the short term, is to show how some of them
can be related to one another and to find ways to include the
others as options.  But I'd be happy to state the goal of
working toward a more unified synthesis (without committing
to a specific deadline for achieving it).

>You seem to be skipping around between the notion that it is best to include
everyone's views on how concepts in an ontology should be defined, and the notion
that inclusion is a kind of methodological first step to faciliate proper "testing"
of which theories work best.

I believe it is a necessary first step.  As we get a better
understanding of the issues, we may be able to discover new
theories that can subsume the older options as special cases.

Whitehead's process ontology, for example, is a very powerful
4D approach, which I suspect may be able to subsume the more
conventional 3D and 4D versions as special cases.  However,
a lot more work is necessary to demonstrate the mappings.
I don't know whether that work can be completed before we
want to release an SUO 1.0 standard.
 
>Those two views are in prima facie tension, at least if you are willing to
admit that the results of any "tests" that would show us how to proceed might
necessitate expunging a particular view (say, on the proper definition of an
"event") if other definitions were demonstrably superior in the context of the
stated purpose.

I think that some special cases may be shown to be undesirable,
others may be related as transformations of one another, and
some may fall into disuse in favor of others.  But I doubt
that every set of incompatible options will disappear.  So I
believe that support for options will have to remain for
a long time.

The remainder of your note (copied below) raises some
other good questions.  Everybody would like to have good
tests and benchmarks that assign a universally recognized
figure of merit to everything from computer performance to
basketball players and corporate CEOs.  I'm moderately
hopeful that moderately useful tests may be developed,
but I'm not holding my breath.

John
______________________________________________________________ 
 
>On the topic of tests:  this is a beguiling idea, I've often thought of what
it would mean to put an upper ontology to some sort of test to measure its usefulness
for its intended aims (i.e. in the SUO scope and purpose documenation), but
it seems to me that the kind of testing that would provide some benchmarks for
fulfilling its stated objectives won't have much to do with disagreements over
how to define things like "event" anyway.  

>Perhaps I'm wrong.  How might we test an SUO to see whether different definitions
of 'event' will make substantive differences in application development areas
and system interoperability issues involving data exchange?  I'm not sure I
know what a test that, say, precluded one but not another philosophical interpretation
of the nature of events when designing (for example) an e-business application
using the SUO would even look like. I'm not even sure this makes sense. How
about anyone else?  
> 
>I know a lot of people have been wondering about much more modest problems
that are nevertheless in the space of ontology considerations, like to what
degree should you axiomatize an ontology based on its intended use, how many
terms and rules it should contain to faciliate target objectives, and so on.
 These considerations are to be considered in contradistinction to, say, the
debate over how to fit both the scientific and the folk psychological conceptualizations
together in a single ontology to make su
> 
>Mike Uschold wrote some  papers on how ontologies can be categorized and tested
to facilitate important automated tasks.  They seem pretty sensible.  Here's
one that's a bit dated but still a nice read: http://delicias.dia.fi.upm.es/WORKSHOP/ECAI98/final-papers/uschold.ps

>
> 
>Erik
>