On Thu, 2004-02-19 at 07:09, Smith, Todd wrote:
> Natural Language processing is very hard to do
> and without some breakthroughs in AI unlikely to produce the results
> requested.

Yes indeed. It is now quite easy to get computers to intelligently
manipulate natural language, but we are still a long, long way from
getting them to really understand it. Witness Doug Lenat's Cyc project,
to model all of common sense (see http://www.cyc.com and of more
interest, its open source subsidiary, at http://opencyc.org ) - many
hold him and his team in awe, but they have been at it for 25 years and
there is still a very long way to go - although there are some
increasingly useful results now appearing, which is why OpenCYC is worth
keeping an eye on.

> A CAD tool is very powerful and very complex but in the right hands can
> produce much of value.  I think that a CAD-like tool can be used be a domain
> expert with some training and still be able to capture the business essence
> that is being lost in the translation from requirements capture and final
> deliverables.

The openEHR people are working on just such tools - to help domain
experts distill their knowledge into Archetypes. It remains to be seen
how successful they will be, and I suspect it will take several
generations of such tools before they are suitable for wide use - at
this stage, we haven't seen the first generation, but I know it is being
actively worked on.

-- 

Tim C

PGP/GnuPG Key 1024D/EAF993D0 available from keyservers everywhere
or at http://members.optushome.com.au/tchur/pubkey.asc
Key fingerprint = 8C22 BF76 33BA B3B5 1D5B  EB37 7891 46A9 EAF9 93D0


Attachment: signature.asc
Description: This is a digitally signed message part

Reply via email to