This is in answer to Brian's claim that random processes can't produce meaningful sentences. I wrote a program which would decode things automatically by using randomly generated sequences. I then looked at the 'decoded' form and selected words and parts of words that appeared from the random decoding. It took about 20 minutes (most of which was spent writing the program for decoding) to generate a randomly generated false decoding of the encoded sentence. For review the original sentence was:
which was correctly encoded as
and then the decoding from random sequences gave the result
Thus it isn't the random mutation that finds the rare sequence, it is the selection that solves the problem in a short time. I didn't plan the sentence above, it was what appeared in the form of whole words or parts of words.
This archive was generated by hypermail 2b29 : Thu Sep 21 2000 - 17:01:59 EDT