Here is
a paper (via
robots.net) on evolutionary programming that gives a good overview of
genetic algs vs.
evolutionary strategies vs. evolutionary programming.
One of my mentors in this industry once told me that he thought evolutionary programming in some form would be needed to tackle the complexity of future applications. These applications would determine the best solution to a problem using evolution and fitness tests that the programmer would create. Would this ease the burden on the programmer? Probably not. But it will allow one to create much more complex applications.
But how do you take this model of programming and turn it into a tool that can be used by many?