What Darwin didn't know.

[Aspman]

Ya ok I cede. Going through this stuff again reminds me why I didn't go for a masters or PhD.

[JewishIntent]

Heh, going through this stuff makes me want to go back to school for sciences. I had taken one year at a 4-year college as a bio/chem major and never went back. Now i work in a corporate HR department. . . . still have no idea how i ended up in this proffesion.

[bballad]

Relyt: your attack on evolution is rather pathetic but others have dealt with that very well already. Your lack of understanding of how computer programming works is appalling. I suggest that you look into Danny Hillis (if any of you don't know that name and consider yourselves computer people be ashamed he is the greatest mind in computer science of our era). You will see that computer programs do evolve randomly through natural selection... in fact that may be the best way to write a program.

EDIT: Figuered I should let you know what evolutionary programming is. Basicly you take a lot of small randomly generated programs and put them in an enviroment to do something....they spawn (programs that come closist to doning what you need them to do get combined) to form the next generation of programs....on and on it gose and a working program that dose waht you need it to is spit out the other end. So yes programs are written by random mutations.

[JewishIntent]

bballad: great way of explaining what evolutionary programming is, but your statement gave rise to a question for me

You will see that computer programs do evolve randomly through natural selection

I'm qurious if we would actually consider natural selection random, since there is a purpose/reason behind what is selected and what is not.

[bballad]

Truthfully you have a situation setup to lead to your solution so teh selcetion is not compleatly random..and in dealign with computers you only ever have pusdorandomniess.


But the seed code is as random as a computer can get. To me it is one of the most conclusive practial proofs that evolutionary theories work.

[ark_templar]

Hey actually on reading the above set of posts, I remembered reading an article in Scientific American about a robot that learnt how to fly and basically its programming was done using evolutionary programming its basic aim was to reach the heighest altitude possible with its wings. It was found adopting the same wing flapping pattern as birds. interesting side note. It also tried "cheating" by "standing" on its "wings". note also that the actuall lift input generated was on the basis of a simulation thus although it did flap it never lifted of the ground it was only "fooled" into believing that it was lifted of the ground.

I am trying to look for this article if I do find it I'll post it as soon as I can.

[Aspman]

There was also some chap who had 'learning' analogue robots. Very simple things with hand soldered boards and just a couple of motors moving bent wire 'legs'. He just built them switched on the battery and set them down, after a while they would 'learn' how to walk and were resilient in that you could cut off one or two legs and it would relearn and adapt. But there was no programming involved it was just very basic electronics.

I'll post a link if I can find anything.

<edit> They're called BEAM robots

http://www.nis.lanl.gov/projects/robot//
http://en.wikipedia.org/wiki/BEAM_robotics