Path: utzoo!mnetor!uunet!lll-winken!lll-tis!ames!pasteur!ucbvax!SUMEX-AIM.STANFORD.EDU!SCHMIDT
From: SCHMIDT@SUMEX-AIM.STANFORD.EDU (Christopher Schmidt)
Newsgroups: comp.sys.xerox
Subject: (none)
Message-ID: <12397384509.22.SCHMIDT@SUMEX-AIM.Stanford.EDU>
Date: 11 May 88 06:27:03 GMT
References: <1581805@ub.cc.umich.edu>
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The Internet
Lines: 31


    Christopher, I would agree that there are a lot of neat ideas in
    the Interlisp-D environment, but I disagree that this is quality
    software. I seen some of the sources (older ones, I'll admit) and
    have to say it was some of the kludgiest stuff I've seen.

I forget who made the observation first, but "the brain is a kludge."
It does arithmetic slowly, has quite a high error rate, and doesn't
do terribly well on any of the Gabriel benchmarks, but I wouldn't
exclude it from my list of quality AI products.

I hope people don't mind my "having the last word" and requesting an
end to this discussion of `quality' on Info-1100.  After reading
dozens of pages of _Zen_and_the_Art_of_Motorcycle_Maintenance_ devoted
exclusively to explaining the writer's notion of `quality' with little
or no increase in my understanding thereof, I doubt we're going to
reach any conclusions in this forum.  A better forum for this discussion
would be AI.WORKS@OHIO-STATE.ARPA, but I doubt any minds will be changed...


If you do retire your d-machines, please seek a less fortunate lab
to give them to.  I'm sure there's an AI lab somewhere that they can't
afford to put Apollos on every desk.  I for one am hoping that someone
out there will donate a disk drive and a set of circuit boards for
us to use as spares for our 2 dorados...   Info-1100 would be a good
place for you to make such an offering.  ("For sale" notices, however,
would constitute commercial use, and can't be permitted on the arpanet.)
	Best wishes,
--Christopher

P.S.  I removed you from the list, as requested.
-------