So i've been reading about minimax polynomial approximations and have found them to be pretty impressive. However, i am confused on exactly how to determine the constants. The first step is supposed be solving for the Chebyshev polynomials as an initial guess. I'm reading wikipedia but i'm a bit confused on how to approximate a function with them. http://en.wikipedia.org/wiki/Chebyshev_polynomials From there I use those factors as an initial guess into a linear system of equations But i'm confused on exactly how to determine the error function (more refined grid?) also how do you iterate through it? Keeping adding dimension until your error criteria is met? Bit lost when I read about it. Lot's of people talking about it but almost no examples. any code that can be linked or any questions that could be cleared up here would be much appreciated.