First page Back Continue Last page Overview Text
Notes:
Since floating point is a complicated world, we want to follow standards.
The IEEE is a professional group for programmers. Among other things, they write standards. IEEE 754 is the standard for binary floating point. 854 is a generalization that covers decimal as well. It looks something like this... Okay, it strains my brain to read this, but that's why the world has experts.
Mike Cowlishaw, of IBM is one of the leading proponants of decimal arithmatic, and he has written a very specific spec for decimal floating point conforming to IEEE854. He includes not only detailed specs but an exhaustive test suite. So rather than inventing from scratch, the Python crew were able to simply implement Cowlishaw's spec, and use his test suite to verify conformance.