From: John Maddock (john_at_[hidden])
Date: 2006-10-31 13:39:36
Paul A Bristow wrote:
>> But we have the option of doing better. I'm wondering if we
>> should change
>> our code to use the better approximating fractions.
> Very probably would be better. But I'm not the maintainer of lexical
> cast code, etc, merely an interefering bystander ;-)
We also know that the current approximation works, and is supported by some
half-decent mathematicians, so maybe we should be conservative? Unless
there are cases where the alternative works better, *and* that the
alternative will never under-estimate the number of digits (overestimation
being benign I assume?)
Just my 2c worth...
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk