Optimal rational approximation

from blog John D. Cook, | ↗ original
A few days ago I wrote about the approximation for cosine due to the Indian astronomer Aryabhata (476–550) and gave this plot of the error. I said that Aryabhata’s approximation is “not quite optimal since the ripples in the error function are not of equal height.” This was an allusion to the equioscillation theorem. Chebyshev […] The post...