On Tue, May 24, 2005 at 12:40:06PM -0700, wj wrote: > > Fixed point is a "lossless" way to represent decimal fractions. Standard > > floating-point uses binary internally, and so "0.01" can't actually be > > represented properly (just as 1/3 can't be represented completely in > > decimal -- 0.33333...). Fixed point is generally implemented by using > > integers and keeping track of where a decimal point ought to be. (Addition > > is straightforward, multiplication/division gets an extra step.) > Fixed point is _not_ lossless. It is exactly what the name is implies, a > floating point format where the decimal point is in a fixed location. I > am very that there will be some "loss" when representing the number 1/3. > You either have to store it as a fraction or there is loss ;-) Hence the quote marks -- I didn't mean that there wasn't any rounding. :) -- Matthew Miller mattdm@xxxxxxxxxx <http://www.mattdm.org/> Boston University Linux ------> <http://linux.bu.edu/> Current office temperature: 73 degrees Fahrenheit.
