Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: Re: SUO: Re: Proposed SUO Content Outline




>I'd appreciate any way to make it easier to navigate through
>this ontology. I don't use Lisp, and would request that a
>neutral format like XML/MathML be used for interchange .

See another recent message from me for comments on XML; basically, 
KIF can be put into XML syntax if you prefer it that way. It is 
exactly the same task as writing a KIF parser.

>The current file is too big, and can be carved up into
>several 'domains of imterest' , linked with hyperlinks.
>Modularity is the only way to proceed here , even
>Cyc uses micro-theories,

That is potentially misleading, however. Microtheories in Cyc's sense 
are not partitions of the concept heirarchy, but of the axioms. The 
Cyc ontology heirarchy is made MORE complex to browse through, if 
anything, by the presence of microtheories, since the same term may 
have different meanings in different microtheories.

> IMPS uses 'little theories'
>(http://imps.mcmaster.ca/theories/theory-library.html)

That is indeed an impressive piece of work, but mathematics is a much 
tidier subject than general ontology design. We can aspire to this 
kind of partitioned elegance, but I doubt if we will find it easy to 
achieve.

>ISO Std 13568 Z uses schemas, XHTML uses modules, C++
>uses classes and so on.
>
>The real advantage is that a Subject Matter Expert for the
>specific domain of interest can initiate and maintain the
>work of defining how that domain communicates its work.
>For example, a website just specifying what a dam on a river is:
>http://www.spatial.maine.edu/~brugg/specs.html
>
>Its a division of labor effect as much as anything.
>
>All I need in the structural ontology is a way to glue
>theories together,

That is rather like saying all you need is a magic wand. True, but :-)

> and a minimum set of descriptors such
>as Actor, State, Object and Interaction!!!

But those terms already presuppose a large set of highly 
controversial ontological issues. You seem to have settled on a 
case-grammar approach to ontology design, but that is quite at odds 
with many of the others being contemplated.

>Just a few v general
>descriptors, plus a notion of refinement to iteratively take one to
>the next layered level of detail.

Can you explain what you mean by a 'layer' here?

Pat Hayes

---------------------------------------------------------------------
IHMC					(850)434 8903   home
40 South Alcaniz St.			(850)202 4416   office
Pensacola,  FL 32501			(850)202 4440   fax
phayes@ai.uwf.edu 
http://www.coginst.uwf.edu/~phayes