* Ingo Menger:
> On 2 Mai, 07:37, "xah...@[EMAIL PROTECTED]
" <xah...@[EMAIL PROTECTED]
> wrote:
>
>> We could, for example, say that all languages in year 2020, will all
>> have full mathematcial functions build in.
>
> What exactly it means for feature x or y to be "built in" or whether
> it makes any difference depends on the language itself. Most of the
> time a user will see no difference anyway.
And the intersting question is not how much code it takes to type B_2n,
but how to implement a system that can (a) infer a bound for the
denominator of B_2n, and (b) can use that knowledge to efficiently
compute B_2n based on a sufficiently exact approximation of
\pi^{-2n}\zeta(2n). If I read the article correctly, Mathematica can't
do that automatically, it's been told by a programmer: "But a few years
ago I programmed a quite different algorithm into Mathematica."
|
