[Courses] [C Doubt]

[Laurel Fan]

On Wed, Oct 09, 2002 at 04:22:27PM -0700, Daniel. wrote:
> So the first integer that can't be stored exactly would be (in binary)
> 10000000....0000001, where there are 52 zeroes? The range for storing 
> integers then goes from 1 up to 2^53, and from -1 to -(2^53)? 

Essentially correct.  You can also represent additional power-of-2
multiples of those integers (constrained by the 11 bits of exponent).

> (No way to store a 0?)

In the scheme I described, there is indeed no way to store a 0.
However, one of the details of the standard that I omitted is that 0
is represented by setting both the fraction and exponent to 0.

-- 
laurel at sdf.lonestar.org
http://dreadnought.gorgorg.org