[johnisgood]

> Repeat it thousands or million times, and you get a far, far better result than what you started with.

I wish.

[cassianoleal]

That's why I never write a stop clause in my recursive functions. Makes everything infinitely better. /s

[johnisgood]

Thank you! I will try that and see how that goes. :)

Edit: I got a stack overflow! Does that mean I should download more RAM, or what am I doing wrong?!

---

But seriously, I think the author is referring to evolutionary algorithms. It does not make much sense regardless. Smith and Trinity are crucial to account for, but the author left them out completely.

[dragonwriter]

> or what am I doing wrong?!

Not using a a language implementation with tail call optimization, or recursing outside of tail position, probably.

[johnisgood]

It better be perfect though!!! At what point can I tell that it is in the most perfect state? If we are talking about infinities, is there a point where we can say that it is in a perfect state and it cannot get any more perfect? Is there such a thing as perfect state?

For example:

> It simply means taking something you have, running it again, but running in such a way that it gets slightly better next run.

So... is there an end of this? Physically speaking, there is a limit, yeah, but can we reach the most perfect state where it cannot get any better?

[cassianoleal]

> is there a point where we can say that it is in a perfect state

I suspect it's asymptotical. You'll need to determine at which point it's close enough to perfection to stop improving, or just let it rip to infinity (and never benefit from it, probably).

[johnisgood]

Yeah, fair enough. I am interested in cases where we do not know what we are looking for exactly, or rather, we do not know exactly what "perfect" is. When we can define the state we want perfectly, then we have an easy job, but what if we cannot?