21 Jan
2007
21 Jan
'07
6:30 a.m.
Henning Thielemann wrote:
Certainly no surprise - there is already an implementation of continued fractions for approximation of roots and transcendent functions by Jan Skibinski: http://darcs.haskell.org/numeric-quest/Fraction.hs
Wow, nice. Now - how was I supposed to have found that? It has been know at least since early Knuth that continued fractions are the best approximation to rationals (in a certain sense). It is too bad that Data.Ratio took the idea of "simplest fraction" from Scheme for the approxRational function. I wrote my own replacement using continued fractions. Now I see that I have reinvented several wheels. Thanks, Yitz