On 6/28/06, Andrews, Mark J. <MarkAndrews_at_creighton.edu> wrote:
> When you say "falls apart," do you mean traversing the ontology takes to
> long and doesn't return useful results, or something else?
All ontology work is based on persistent identity of entities, there's
absolutely no way around that, but establishing identity is a very
human thing ; it is opinion on what something is. In a computer system
we need to replicate these opinions in a coherent way, otherwise
you'll be tied to closed domains, which is pretty much where we are
now because *that's* where they work great. In a SemWeb scenario,
identity needs to be the *main* mechanism for inference, yet all we
currently have is the fuzzy notion of a URL. It simply isn't good
enough for serious work.
Now with Cyc, it too works fantastic in a closed domain, but in trying
to apply the enteties defined in Cyc to a diverse set of systems that
needs to share those identities, and there is no mechanism for doing
so.
Not sure this makes much sense to most people, but there is a huge
problem with identity definitions for computers as the *human* notion
in this case doesn't translate well to them. I can ellaborate more,
but I feel this is rather off-topic. :)
regards,
Alex
--
"Ultimately, all things are known because you want to believe you know."
- Frank Herbert
__ http://shelter.nu/ __________________________________________________
Received on Wed Jun 28 2006 - 19:52:18 EDT