Fixed Point Problems

1)
I assume these are the three base 10 representations expected:
two's complement: -9855
one's complement: -9854
signed magnitude: -2433
unsigned binary: 55681
And for a little added value:
excess/bias: really anything one wants, given the right bias


2)


3)
	in 2s complement
	1) cec9178d
	2) fffe4000
	3) 00010000
	4) fff8c000

Floating Point Problems

1)
	in hex:
	1) 43806000
	2) 3f6872b0
	3) c5801900
	4) 332bcc77
	5) c2800000

2)
	in hex:
	1) bfe00000
	2) 3f7fffff
	3) c0e80000

Extreme Issues

1)
	It's unclear why zero must be a "signal" value.  I see no problems 
resulting from simply classifying it as a denormalized number.  In which 
case it would be 0 x 2^-126 = 0.

2)
	127
	in hex: fe

3)
	-126
	in hex: 01

4)
	Two examples are the smallest and largest postive, denormalized 
numbers:
	in hex:
	00000001
	007fffff