- From: Drew McDermott <drew.mcdermott@yale.edu>
- Date: Tue, 16 Dec 2003 20:25:22 -0500 (EST)
- To: public-sws-ig@w3.org
[Mike Huhns]
Well, here is a more difficult example to add to Drew's list. My
project involves the simulation and visualization of a human heart. One
ontology is based on a surface model of the domain (a heart is
represented as a collection of small flat polygons) and another ontology
is based on a volumetric model (a heart is a collection of voxels, i.e.,
small volumes). Translating back and forth is a mathematical, not
linguistic problem (requires partial differential equations). Is that
still straightforward deduction?
No! On the other hand, it's not as hard as, or hard in the same way
as, translating natural language.
That's a good example, Mike. It points out that in some cases any
translator or merger is going to have to call some procedural module.
Which of course won't magically solve all the problems. One can
produce a black box that translates polygons to voxels, but it won't
be able to do anything with a statement that mixes existential
quantifiers in with the descriptions of constant polygons.
-- Drew
--
-- Drew McDermott
Yale Computer Science Department
Received on Tuesday, 16 December 2003 20:26:33 UTC