Why does D not have generics?

Sat Jan 16 08:41:18 UTC 2021

On Friday, 15 January 2021 at 18:11:04 UTC, Bruce Carneal wrote:
> value types of the same size.  This shows up when implementing 
> radix sort where one solution is to map to/from whole numbers 
> (NaN semantics being ignored).

If we accept the following encoding for floating point then it 
should work out ok?

INPUT               OUTPUT
s  exp  sig         s  exp  sig

1  1100 1111110     0  0011 0000001
1  0011 0000001     0  1100 1111110
1  0000 0000000     0  1111 1111111

0  0000 0000000     1  0000 1111111
0  0011 0000011     1  0011 0000001
0  1100 1111110     1  1100 1111110

Algorithm:
1. flip the signbit
2. if the flipped sign is 0 then negate the exponent and 
signficand/mantissa.

And it can be reversed just as easily?

Not too bad!?