First page Back Continue Last page Overview Text

Notes:


Now if someone is going to create a floating point system, there are lots of design choices they need to make.
There's basic stuff like [...].
For instance, the floating point that is built into your CPU uses a radix of 2, while “Decimal” floating point has a radix of 10. This is the most fundamental difference and is the reason for the disagreement we saw earlier.
There are also a lot of fancy details to work out [...]. These kinds of choices all interact in complicated ways.
Fortunately Facundo and Tim Peters happen to be a couple of the world's greatest experts on Floating Point numbers. Actually, they're not. But we DID employ the services of some of the world's greatest experts. How?