You are on page 1of 1

Evans,

 
I like how you organized your thoughts into two distinct focus
areas.  First, you talked about "reasons to store the floating-
point numbers in normalized form". Then you explained the
advantage of using bias.  With that, I would offer that
there are also situations where you would use floating-point
data type to represent integers.  They include:
 
 A situation involving JavaScript.  In theory, it does not

have integers; and only supports floating numbers (i.e.,


doubles).
 
 Another situation involves computer graphics.  A

programmer does not need to have exact accuracy in 3D


graphics.  It just needs to "look right."  A programmer will
have many object calculations in 3D space.  Things like
rotating or scaling requires floating point calculations.
The floats are supported by graphics hardware.  They are a
good choice in the domain.
 
Finally, I also noticed that you pointed out that a "computer
can't represent negative integers, so by using this method we
can represent the negative integers".  This understanding
underscores very clear and concise insight.
 
Awesome.
 
Best,
Dan

You might also like