fixed point vs fixed decimal

Terry Hancock hancock at anansispaceworks.com
Wed Jan 8 02:41:35 CET 2003


John Roth wrote:
> You're right! I'd forgotten about the applications in graphics and
> motor control where floats are too expensive.

They are also a convenient natural representation for angles -- if you 
allow a fixed size, unsigned integer to overflow without error, and map the 
range of the integer to a full circle, then its behavior is exactly correct 
for the task and does not require special edge-case handling.

I saw this trick in a graphics programming book years ago.

They can also be handy in graphics acceleration tasks (but in C, obviously).

-- 
Anansi Spaceworks
http://www.anansispaceworks.com




More information about the Python-list mailing list