Free Republic
Browse · Search
News/Activism
Topics · Post Article

To: gore3000
You cannot just insert what is essentially new code in the middle of a program at random.

The procedural programming analogy only goes a little ways. The genetic/biological system is more like an event-driven object oriented system. You can plunk down whole libraries of new object classes into a program with no effect at all.

Plus, the way an object method gets invoked is completely different. In a computer program, you have to get the name of the method (or its starting address if we're talking machine code) exactly correct or else the compiler/runtime gives you an error & the whole thing stops. It's all or nothing with traditional programming languages.

If I were to simulate a genetic system, it'd be more like: the probability that an object's method gets invoked is proportional to how correctly it was spelled in the statement that called it. That's just for starters. A totally different programming paradigm.

I've never tried to write a GA program or even seen a genetic programming language, so for all I know they already have languages like that. If they do exist, I'm sure you could duplicate objects & methods and mutate them & have them sometimes end up doing something useful.

47 posted on 08/11/2002 9:00:51 PM PDT by jennyp
[ Post Reply | Private Reply | To 44 | View Replies ]


To: jennyp
The procedural programming analogy only goes a little ways. The genetic/biological system is more like an event-driven object oriented system. You can plunk down whole libraries of new object classes into a program with no effect at all.

Plus, the way an object method gets invoked is completely different. In a computer program, you have to get the name of the method (or its starting address if we're talking machine code) exactly correct or else the compiler/runtime gives you an error & the whole thing stops. It's all or nothing with traditional programming languages.

If I were to simulate a genetic system, it'd be more like: the probability that an object's method gets invoked is proportional to how correctly it was spelled in the statement that called it. That's just for starters. A totally different programming paradigm.

Well, you have a problem there with chance mutations already don't you? Wrong spelling. To get correct spelling by chance seems an unreasonable assumption.

However, the big problem with your explanation is that all programs are in machine language. What you call object classes are just for ease of writing programs. When turned into executables, they are all machine language programs and cannot be changed by random insertions of code.

Also you need to realize that any program needs to differentiate between data and code, in the case of an organism between the gene data and the DNA code to make them work. It requires intelligent interpretation of the DNA 'bits' to accomplish this. More importantly though, what this shows ( the presence of both data and code) is that you cannot just change a gene, add a gene, to get new functioning, but you also have to simultaneously change the code. This is a bit much to expect from random mutations.

52 posted on 08/11/2002 11:03:47 PM PDT by gore3000
[ Post Reply | Private Reply | To 47 | View Replies ]

Free Republic
Browse · Search
News/Activism
Topics · Post Article


FreeRepublic, LLC, PO BOX 9771, FRESNO, CA 93794
FreeRepublic.com is powered by software copyright 2000-2008 John Robinson