The thing is, often after a great answer gets posted, someone will come along and edit the question into a better shape (for the sake of it not getting closed). So, it's hard to find some of them.
but for math, you know absolutely for certain that it's homework, since homework questions and real-world questions look different
of course, it's better than the old way
in the earlier days of the internet, people would refuse to help someone or even link to an answer sometimes, if they knew that an answer existed that the asker could potentially find for himself
@Eric By obvious, I meant obvious. On SO, this can look like "You are given n balls and x cups.... write an algorithm to describe... make sure you use coding practices as established in..." :D
@Optimizer you don't understand; you have a network of neurons, and the input to it would be cat pictures. Typically, you train a network on a rather large set, though
This is not really the kind of question that succeeds around here, though. For one, building a neural network is a lot of work, and training it to recognize photos is extremely hard
training it to recognize English letters for OCR is already hard
@TAbraham When you say "find a cat face", does that mean simply detecting if a cat face is present, or outputting the coordinates of things like eyes/nose/etc?
Ok, can I ask about rules 1&2? Normally built-in libraries are discouraged for code golf, since it then boils down to "what language has the best built-in function". Also, not counting whitespace is unusual. Any particular reason?
@Geobits well, because this project wouldn't be possible without using built-in functions.. Oh, if you want, I can count whitespace if that is what genrally people do..
@TAbraham Anything's possible. I was just noting the direct opposition to the normal rule of "no built-ins" here. It seemed unusual enough to ask about. Maybe you should say you may use them instead of may only?
Well, you should probably just let me think about it a little more, and then ask around. If you don't have any progress for i=2, it's probably going to come off as rude
@EricTressler thanks! I suspected at one point that someone who knew how to reformulate the problem in terms of polynomials and could do some magic discrete Fourier analysis might be good
there is David Speyer
but maybe I only think that would help because it is outside my expertise so it looks really clever to me :)
also consider unrolling your vector and thinking of it as an infinite repeating vector of period n
and applying what he does
@Optimizer back to the palette thing: thanks, I'm glad you like the current scheme. I wanted to add a fractal toy subpage, though, and I wanted it to be kind of organic-looking. That's what the new scheme was supposed to be for
I'll ask about it again after I've built something, so you can see it in context