home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!charon.amdahl.com!pacbell.com!sgiblab!sdd.hp.com!usc!news.service.uci.edu!ucivax!ofa123!David.Rice
- From: David.Rice@ofa123.fidonet.org
- Newsgroups: talk.origins
- Subject: Random Chess Moves
- X-Sender: newtout 0.02 Nov 17 1992
- Message-ID: <n0ea4t@ofa123.fidonet.org>
- Date: 17 Nov 92 09:03:16
- Lines: 58
-
- Who: amorgan@Xenon.Stanford.EDU (Crunchy Frog)
- ID: 1992Nov13.224645.12836@csd-newshost.stanford.edu
-
- [mid-sized cut]
-
- DR> "There -ARE- computer programs that introduce randomization of
- DR> its internal code, to produce better code. I've read of a chess
- DR> program that does this! There are also defence department wargame
- DR> computers that randomize, then retain the successes and discard
- DR> the failures, so that "better and better" routines are selected
- DR> for.
-
- CF> "Another point is that the Tunster has been stipulating that
- CF> we use well designed, carefully optimized code as our chess
- CF> program. Funny, most animals I see in the world wouldn't
- CF> qualify as "well designed" by a mile. If you want to make
- CF> a computer program analogy, take a bug ridden, hacked up
- CF> checkers program written in C that a Pascal programmer converted
- CF> to play chess in a long weekend. Then see if you can make
- CF> *that* better by random changes. A perfectly designed program
- CF> almost by definition can't be improved, neither can a perfectly
- CF> evolved animal. Neither one, however, exist."
-
- Crunchy is right--- a very complex system, or several systems
- working together, lends itself to psudo-random selection (reward
- and punishment), while simple systems do not do well with random
- attempts towards a goal (goal for life being "to survive long
- enough to reproduce, and secondly to survive long enough to
- defend that reproduction).
-
- My twin brother works in telecommunications (poor bastard!). He
- told me about the programs California TelCo use to find the best
- route from point "A" to point "B" ("best" being the least number
- of physical connections, or distance, or cable medium). Now, one
- would think they'd do it logically (SMILE!), but some of their
- programs use trial and error, within programmed constraints.
-
- Once a path is determined, the program remembers to use that path
- in other searches, yet it also -strives to improve- that path in
- background time slices.
-
- As it stands now, telecommunications is a sprawling monster that
- covers almost all land masses, and most are connected in some way
- to each other. Throw in communication relays that are in orbit,
- radio and microwave stations. . . . telecommunications EVOLVED
- and looks designed (hee! Sheeeeit!), but it has just as many
- kludges as living organisms. No one really understands Ma Bell.
-
- CF> If you are going to make stupid demands Lionel, you could at
- CF> least stick to the point.
-
- And my point? Er, ah, no point. Computers are great at compressing
- hundreds of millions of years of evolution via natural selection
- into the span of one's lunch break during work. Can you imagine
- how much better life forms would have been designed if God had
- had a computer?! (GRIN)
-
- CF> C Frog
-