The first should (for this method to work) be fitter than the second, but I don't know a way to quantify that.

Also, p(syntax error) is very high for constructing a valid Perl string randomly. The chances are that all members of a population would generate syntax errors - at which point they'd all have the same fitness.

A better approach would be to create a baby language safe (or safer) from syntax errors and eval that. Like maybe just arithmetic operators and numbers.

Although - it's just occured to me - you'd be on fairly safe ground with a subset of a highly structured language like Logo. And you could evaluate fitness by checking how similar the figure drawn was to your target figure.

Andy.

p is not a syntax error (test it) . Or do you mean a string which is a valid program and contains the chars "syntax error"? Then you have misunderstood. I meant "syntax error" in $@ after eval

In Re (Obfu generator for): Ode for getprotobyname, I randomly generate tr/// patterns, and check if the "just another perl hacker" message can be generated as a result of that transliteration. Perhaps you could try generating such patterns genetically, with the goal function being the number of different characters you cannot generate.