float / double support in Python?

[Paul F. Dubois]

To answer your question directly, which others seem not to have done:

Numerical Python (numpy.sf.net) supports many types, including bytes, 32 bit
floats, etc.

Numerical Python is widely used. A successor product 'numarray' that will be
easier to maintain and more robust is being worked on but it will be a
mostly-compatible replacement, not a bifurcation of the community. Do not be
confused by uninformed messages on this topic.

One hope is to make numarray eventually part of the standard library.

Part of Numeric is a module 'kinds' (PEP-242) that offers some support for
platform-independent floating point precision. It could be developed further
to include a 32-bit float scalar type.