Of course Python has no chance of getting this right. As Feynman says, you have to multiply 2*pi by 10^100, and throw away the integer part; so the right place to start is a record of pi to 100 decimal places.
(I really need to be more careful; this isn’t the first time I’ve been caught making a trivial error in public, and it’s starting to get embarrassing.)
Of course Python has no chance of getting this right. As Feynman says, you have to multiply 2*pi by 10^100, and throw away the integer part; so the right place to start is a record of pi to 100 decimal places.
(I really need to be more careful; this isn’t the first time I’ve been caught making a trivial error in public, and it’s starting to get embarrassing.)
Well, what I wrote was wrong too—what matters is the non-integer part of 10^100/(2*pi).