You are going to block this site. This will do the following:
- You will no longer see this site in searches.
- Site will no longer see your site in searches.
- Site will not be able to comment on your site profile.
- Any comments this site has posted to your profile will not be displayed.
Are you sure you want to do this?
"Almost any" means anything except some crazy things. At least if the graph is not "discontinuous at infinitely many points in a period", you can use Fourier series without worry. It's because periodic functions make a function space, and sines, cosines make a coordinate system for the space. It's like (x,y) = x(1,0) + y(0,1). Instead of (1,0) and (0,1), you just use sines and cosines.
You should use infinitely many sines and cosines because obviously the dimensions of those functional spaces are infinite. But you can approximate with sufficiently many sines and cosines anyway.