rpw3@fortune.UUCP (12/16/83)
#R:seismo:-47200:fortune:8600005:000:1527 fortune!rpw3 Dec 15 23:50:00 1983 And I keep telling people that our [xyz proprietary feature] measurement numbers are hogwash! No matter how educated or experienced in UNIX a person is, if they don't really know (gut level) the hazards and pitfalls of experiment design, and unless they really understand statistics, "Sparse data in ==> Garbage out." Or, just try and get a manager to understand that just because the bug crashed the system 3 times today, and since your swag (sophisticated wild-assed guess) is that it will take looking at 6 more crash dumps to find it, that you will not necessarily have fixed this show-stopper bug in 6 / 3 = 2 days at 4:00 p.m. (It's 4:00pm now, isn't it?) Or that maybe even (for a smarter manager... I mean, after all, I'M a manager), Erlang-B won't solve all of your configuration and sizing problems. (Just how much swap space DOES this system need?) The vast power trigger-induction holds over us hints at something in our evolution (oops!) that says that a few people surviving by making a correct wild guess (and many many dying from the wrong wild guesses) is better than the whole population dying from overconservatism. But since we no longer shoot people who draw silly conclusions from insufficient data, I suppose we had better lend out our statistics books. Me, I never liked the stuff, but I remember enough to win at Backgammon... Rob Warnock UUCP: {sri-unix,amd70,hpda,harpo,ihnp4,allegra}!fortune!rpw3 DDD: (415)595-8444 USPS: Fortune Systems Corp, 101 Twin Dolphins Drive, Redwood City, CA 94065
MJackson.Wbst@PARC-MAXC.ARPA (12/22/83)
"The vast power trigger-induction holds over us hints at something in our evolution (oops!) that says that a few people surviving by making a correct wild guess (and many many dying from the wrong wild guesses) is better than the whole population dying from overconservatism." Most human reasoning employs faulty logic ("most A is B, most B is C, hence most A is C") because in the real environment matters are seldom sufficiently clear-cut (or information sufficiently complete) to enable one to construct formally correct syllogisms. Hence the use of such slippery tools as analogy. A paper by Marvin Minsky, "Jokes and the Cognitive Unconscious," discusses the role of humor in instructing/exercising the mental "censors" which attempt to check clearly inappropriate use of such uncertain but useful mental processes. Mark