accuracy of CLA math

Hi Eric,

for the sin and cosine, the algorithm basically finds the closest match in the lookup table, subtracts the table angle from the original to get the residue and then does a taylor series expansion on that residue - im not sure how to put an upper bound on the accuracy for this but i usually compare it to what MATLAB gives me and it seems to be accurate to within 10^-6 error.

The same goes for the arctan , arcsine and arccosine algorithms as well however those are polynomial fit algorithms and not Taylor series expansion