Hi Mark,
I was a bit surprised to find the very slow Farey approximation by
means of the class in the mxNumber package. If the goal
was to reconstruct a rational from a float it is not a good choice and
should be replaced by a continued fractions approximation. Some time
ago I implemented it by m
Kay Schluehr wrote:
> Hi Marc,
>
> I was a bit surprised to find the very slow Farey approximation by
> means of the class in the mxNumber package. If the goal
> was to reconstruct a rational from a float it is not a good choice and
> should be replaced by a continued fractions approximation.
T
