On 2/26/2011 6:32 PM, Obrst, Leo J. wrote:
> I agree with Jans and John about the usefulness of Prolog and logic
>programming
> more generally. Apologies for the length of this message. (01)
It's a great message. Please don't apologize. If somebody isn't
interested, it only
takes one click to delete a message of any length. (02)
The point that I have tried to make many times is that there has been
half a century
of extremely active R & D in the AI community. Some of it has been
absorbed by
mainstream IT, but much more has been implemented in research projects and
small applications. (03)
Prolog, however, has been used in some huge mission-critical applications.
One of the biggest is Experian. They use Prolog to evaluate everybody's
credit
rating, and they use it so much that they bought Prologia -- the company
that
was founded by Alain Colmerauer, the inventor of Prolog. However, Experian
doesn't advertise their rules -- they keep a low profile. (04)
When I complain about the Semantic Web, I have no quarrels with any
specific project
that anybody has found useful. But what I deplore is the huge diversity
of systems that
have been eclipsed by the tiny trio of SW tools: RDF, OWL, and SPARQL.
The layer
cake has a box labeled RIF, but no implementations to support it. That
makes OWL,
a very highly specialized tool, the only reasoning system available for
the SW. (05)
> We [Mitre] developed our own OWL+SWRL+Prolog tool during research in 2004-2007
> i.e., translating OWL ontologies, SWRL rules into Prolog, and developing a
>composite
> Prolog interpreter/compiler for more efficient runtime reasoning. We had to
>develop our
> own back then because there were no such implemented systems. Now, of course
> there are, including some hybrid systems and SILK. (06)
Yes. But Mitre has many highly talented PhDs and the ability to get
funding from
gov't agencies to pay them. Few universities and even fewer businesses
have the
talent and funding to do that. (07)
In 1998, I was very hopeful that the Semantic Web project would
encourage a thousand
AI flowers to bloom. But today, the only things left in the SW are
three very specialized
posies that have choked out every other flower in the garden. (08)
A couple of weeks ago, Len Yabloko asked whether Watson was good for the SW,
because it caused IBM to shift attention *away from* SW technology
(i.e., the three
posies left in the garden) and move to a multi-paradigm technology that
used a variety
of different AI technologies (mostly oldies, but goodies). (09)
But I think it's great!!!! It demonstrates that those three little
posies are insufficient
to support the full range of AI technology that is needed for advanced
applications. (010)
> We also used various knowledge compilation techniques... But the latter
>techniques
> are used in many Prologs and especially deductive database systems. Probably
>the
> best latter such that we¡¯ve found is the HighFleet (formerly Ontology Works)
>system,
> but we know that Cyc¡¯s reasoning also includes many such techniques, for a
>couple
> of examples. Knowledge compilation is a vastly underserved topic of research
> in recent years, in my estimation. (011)
Absolutely!!! I keep citing that little paper from 1998 by Bill
Andersen et al. on
knowledge compilation. For citations and a few more comments about it,
see pp. 5-6 of (012)
http://www.jfsowa.com/pubs/fflogic.pdf
Fads and Fallacies About Logic (013)
That 1998 paper was the founding idea for OntologyWorks (which changed their
name to HighFleet for customers who can't spell 'ontology'). That proved so
successful that they have more business than they can handle. (014)
We also use knowledge compilation techniques at VivoMind. But none of those
ideas have made their way into the W3C. They're potentially revolutionary
technology, but funding agencies get distracted by the SW hype machine. (015)
By the way, I am not against the Semantic Web. But I hope that they will
open up their technology to a much, much wider range of paradigms. (016)
John (017)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (018)
|