Abstract: In methods of function genration, DDA and SFT (SAITA function generation) nave quite different forms, although their basic characteristics and errors of function generatio,n have no major differences. The errors between a curve generated by differential algorithm and its original given function consists of basic error and stepping error. A function generated by a differential algorithm approximates a function (properly called an approximate 'function) which deviates from its original function. This error is called "basic error". The approximated function can be derived from the given function itself. Curves generated by differential algorithm may also more or less deviate from their approximated function owing to different stepping approaches. This causes "stepping error". Analysis for polynomial functions of the second and third order shows that deviation of approximatd functions from their given functions may be quite significant. The stepping error is generally much smaller than that of basic errors.