Mailing List Archive: 49091 messages
  • Home
  • Script library
  • AltME Archive
  • Mailing list
  • Articles Index
  • Site search
 

[REBOL] Re: RRGib - Random Rebol Gibberish

From: steve:shireman:semaxwireless at: 31-May-2001 5:06

Anton wrote:
> Hello, > > I am writing a program to generate > random pieces of rebol code, (rebol > gibberish,) and post them automatically > to this list for appraisal. > The name for this program is RRGib. > > Here is the core of its functionality: > > forever [print random/only first system/words] > > The ultimate aim is to create self- > sustaining rebol programs - rebol life. > I've got it to the stage where the > program can almost generate itself, > possibly. > Just a few more words and it's done. > > But I need your help in this effort. > All I need you to do is to point out > for me where all the mistakes are, so > that any given random program is more > likely to work.
If you design automatic evaluation (selection) of your random scripts, mate the better random programs with each other, then you have essentially done the Genetic Programming method (Koza 1990) This could be much faster than human evalution of the random scripts on this list. ;-) Rebol is uniquely suited for Genetic Programming, because it almost never crashes bad with wayward code, (and can be tested with try, anyway) Essentially Genetic Programming is a search for a solution, and uses inductive logic to reach a final design, which runs counter-intuitive to most programmer types who are trained to try to use deductive logic. Anywho, thanks for the one liner. Mutate fast, Steve Shireman Your brain may no longer be the boss Firesign Theatre