This will be very usefull also for computing some astronomical numbers,
There is also a test version (0.41test01) that seems to work
on AMD64 linux machines, although there are probably many problems to find.
This version keeps floats at 32 bits and simply corrects bugs arising
from the new sizes; I hope also to support running Pd using 64-bit numbers;
this will probably always run slower but will work better for some kinds of
applications (representing phone numbers, for instance).