Assuming that you are required to make an 8 bit floating poi

Assuming that you are required to make an 8 bit floating point representation, with a 3 bit exponent using excess-3 notation and a 4 bit mantissa using normalized notation. How would the number 0.5 be stored in the byte?

Solution

we will split 8 bits as follows:

1st bit will denote sign bit (+/-)

Next three bits are for exponential bits

Next four bits are fraction part.

So first we will convert 0.5 to binary, then arrange in 8 bit floating point representation.

sign bit is 0, since it is positive
exponential part (excess e part) is 000 (since no decimal)
fraction part is 1000

Final floating point representation is...

0 | 000 | 1000