On 1/26/12, Stavros Macrakis <macrakis at alum.mit.edu> wrote:
> The 'cf' (continued fraction) documentation should mention that cf
> expansions are exact for rational numbers, but are cut off at ratepsilon
> (for floats) and 10^-fpprec (for bigfloats).
I;ve revised the cf documentation to take that into account.
What do you think?
best
Robert Dodier
PS.
-- Function: cf (<expr>)
Computes a continued fraction approximation. <expr> is an
expression comprising continued fractions, square roots of
integers, and literal real numbers (integers, rational numbers,
ordinary floats, and bigfloats). `cf' computes exact expansions
for rational numbers, but expansions are truncated at `ratepsilon'
for ordinary floats and `10^(-fpprec)' for bigfloats.
etc etc. (same as before)