On 10/25/2011 7:11 PM, Obrst, Leo J. wrote:
> The division between Neats and Scruffies goes back in AI to at least the
> late 1970s-early 1980s and probably before. (01)
The division was there from the very beginning of AI, and it is
related to the similar division in comp. sci. The actual terms
"scruffy vs. neat" were coined by Roger Schank who proudly called
himself a scruffy as opposed to people like John McCarthy and
Nils Nilsson who were neats. (02)
John McCarthy (who died yesterday) was always a neat. Marvin Minsky
(less than a month older than McCarthy) was trained as a mathematician
and proved some theorems, but he tended to be scruffier. (03)
Leo
> most of us working in ontologies these days are both, with a predilection
> to one of those poles. I myself am a scruffy neat. (04)
I agree. The lines are very blurred. My major complaint about the
Semantic Web is that it has the worst of both worlds: a language
designed by neats (OWL) that is used by scruffies who would be much
happier (and probably more productive) with a scruffy language. (05)
The Wikipedia article on the distinction has some facts mixed
with a lot of confusion: (06)
http://en.wikipedia.org/wiki/Neats_vs._scruffies (07)
Following is a confused list from Wikipedia: (08)
> Neats
>
> John McCarthy
> Allen Newell
> Herbert Simon
> Edward Feigenbaum
> Robert Kowalski
> Judea Pearl
>
> Scruffies
>
> Rodney Brooks
> Terry Winograd
> Marvin Minsky
> Roger Schank
> Doug Lenat (09)
Putting McCarthy into the neat camp and Schank in the scruffy camp
is beyond a doubt. But Feigenbaum and Lenat jointly proposed Cyc
in the early 1980s, and I would call Feigenbaum the scruffier of
the two. (010)
Lenat's dissertation topic was the Automated Mathematician (AM)
which set out to hypothesize and prove mathematical theorems.
For Cyc, he started out rather scruffy with a frame-based approach,
but he evolved to a very strong logic-based approach over time.
I have some criticisms about Cyc, but they're not about it's
being too scruffy. (011)
Terry Winograd is another difficult to classify person. His
SHRDLU system was written on top of MicroPlanner, which was
loosely logic-based. It could be considered an early kind
of Prolog without unification. And SHRDLU's way of processing
English questions and commands could easily be formalized in
logic. (012)
In the mid-1980s, Winograd wrote a good book on syntax, which
was as formal as anything at the time. But when he started
to write a second book on semantics, he became disillusioned
with the entire field of AI. He became a very strong critic
of AI -- but he was equally critical of logic and heuristics.
He left AI to do human factors. (013)
John (014)
_________________________________________________________________
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 (015)
|