A program written by Tom Schneider's has been credited with producing information by chance and an equivalence of natural selection. He claims to have solved the information problem of evolution. Since I am an experienced programer and knowledgeable  in Pascal, the programing language it is written in; I decided to analyze his claim from this perspective. 

A copy  of the program can be found at: http://www.lecb.ncifcrf.gov/~toms/delila/ev.html

 

1. Dos Schneider's program actually generate information?

 

If the answer to this question is no, then his program is a failure. End of discussion.

 

In his paper Schneider  Starts with Claude Shannon definition of information as "a decrease in the uncertainty of a receiver." ( Schneider's words ) He precedes to conclude that; for molecular systems; there is a relationship between uncertainty and entropy. He seems to be concluding that information for molecular systems is simply a reduction in entropy.

 

Base on this definition, yes Schneider's program does generate information, however I would contend that his definition is too broad. The problem with this definition it that it does not require a code or specific meaning to qualify as information.  In the program most of the simulated genes do carry no instructions of any kind, they have no meaningful code. The only purpose they seem to serve is being scanned for mistakes. The part of the genes that does have a code and  meaning are not placed there by the results of the random number generator, but are encoded by the program itself.

 

Claude Shannon's paper "A Mathematical Theory of Communication" shows the necessity of a coding system and meaning in transmitting information. http://cm.bell-labs.com/cm/ms/what/shannonday/paper.html

 

 The conclusion here is that since the randomly generated portions of Schneider's program lack the presence of any encoded instructions or knowledge, it fails to actually qualify as information.

 

2. The program has a predefined goal.

The theoretical goal of the program is to evolve the information at the simulated binding sites(Rs) to equal that needed to find the sites.(Rf) The practical goal of the program is to get an organism that produces 0 mistakes when the genes are scanned.

 

3. To simulate the theoretical evolution of information, the selection process must be blind.

Natural selection is a blind possess, in that it does not have a preset goal towards which it is working. The program selection possess fail this test. After scanning the gene and adding up the mistakes, the half of the population with the most mistakes is kill off and replaced with the children of the those with the fewest mistakes. Therefore the selection process is selecting specifically for the fewest mistakes, it is there for not a blind process.

 

4. The result of the program come from a  well calculated, preprogrammed push towards the desired goal.

This fact is demonstrated by watching movie of the out put. http://www.lecb.ncifcrf.gov/~toms/paper/ev/movie/index.html.

Since Rf is set at 4 bits, if you watch the one movies that go 1 generation per frame, you see that if the number of bits goes above 4 is get pulled back down to 4bits. This fact demonstrates the results are the product of a preprogrammed push rather than a real generation of information.

 

While it is an interesting program, it proves nothing about evolution, or it's needed random generation of information. Even if by some definition of information it, succeeded in generation information, it is not the result of the ransom processes and blind selection needed to solve evolution's information problem.

 

 

--- Charles Creager Jr