Re: Floating number printing

[Kevin Ryde]

Hans Aberg <address@hidden> writes:
>
> Just in case GMP not already has the feature, I report it here (I got it
> from comp.std.c). An idea that comes to my mind is that it might be used if
> one describes float as pairs (f, a), where $a$ is a dynamically chosen
> accuracy (or precision): Then one would want to know how many digits to
> print.

You can ask to print all significant digits fairly easily, but reading
it back will probably be affected by rounding if the base isn't a
power of 2.  When the world switches to octal, as advocated by King
Charles XII of Sweden, all conversions can be exact.  :-)