Yeah! Embrace the clickbait!
Anyway, I made a quine, a program that returns its own source code as a result of running it, that can change itself. It's sort of a response to this post:
by am_Unition (also maybe mk if you are interested in this). Betcha you didn't expect me returning with something :].
So what encodes the encoders? The random seed is set by the current chain of nucleotides. Then a type of mutation (add or remove a nucleotide) is selected and then it comes out as a changed source code which will now contain a mutated chain to set the random seed to a different value. Copy the source code, run it again and it will do another mutation in the same 'self-encoded' random manner.
Take that, RNA! You are just a more sophisticated version of what I hammered out in barely an hour! :P
And to think, I felt stupid before I got a notification.
Nice tho.
Edit: Devac I will open this and mess with it soon, but my python IDE takes about one hour to "check for updates". Plus, I'm writing in another language right now, and I don't wanna cross the streams :/. Do you ever get in over your head in your research commitments?
- And to think, I felt stupid before I got a notification.
Come on, man. The title is a literal click bait and I knew from the start that it must be doable. You'll see yourself how primitive is this thing. Plus you shouldn't feel stupid, period.
- Do you ever get in over your head in your research commitments?
Constantly.
Let me help you with your fishing technique. I would've said:
- PHYSICIST WRANGLES A MUTATING PYTHON AND CLAIMS VICTORY AFTER AN HOUR'S TIME
I checked out the code briefly. Been almost a year with no Python, I think it'll show. I see some even and odd handling, random number generation and handling, looks like it iteratively renames the file a variant of ACfileGT (the nucleotides), where the capital letters change. Something like that? Edit: lol, you described as much, I think, I missed the fact that the current naming serves a crytographic seed
What I had in mind was a quine in which the actual commanding (the guts of it) randomly, iteratively changed inside of the program, but it would almost always still mutate in a way that was replicable. Maybe it could eventually produce randomized, unforeseen fatal errors (digital cancer) after some large number of repetitions. Still a very open-ended problem, but I think I got a bit over ambitious with it.
- looks like it iteratively renames the file a variant of ACfileGT (the nucleotides), where the capital letters change
This line? No, it just generates all possible 3-element combinations with repetitions from a set "ACGT". There's no file handling as of yet.
- What I had in mind was a quine in which the actual commanding (the guts of it) randomly, iteratively changed inside of the program, but it would almost always still mutate in a way that was replicable.
That's still doable, I think. Just far less likely to produce an actual line of succession. I think that I can do sort of a compromise between my lazy solution and yours "this might have to reinvent compilers" formulation: what about having a procedurally changed block or mutation parameters like 'add a nested conditional determining a future mutation type' or 'remove this loop that does something the nucleotide chain'? That only needs to comply with Python's syntax but you can add pretty much anything.
Anyway, I told you that it wasn't impressive. :P
While searching for help with a module error I got invited to Google foobar, which sounds like something that will keep me distracted for a while.