1

IEEE 754 stored exponent using Excess K forma, which is also known as Offset Binary representation. I've googled and it seems that it's just one of formats to represent negative binary numbers. I'm wondering why two's complement form is not used for that purpose?

Daniel A. White
  • 187,200
  • 47
  • 362
  • 445
Max Koretskyi
  • 101,079
  • 60
  • 333
  • 488
  • I'm voting to close this question as off-topic because it is not about a programming problem. – Pointy May 07 '16 at 12:05
  • where do you suggest to migrate it to? – Max Koretskyi May 07 '16 at 12:08
  • I'm not sure; it's a question about decisions made by a committee 40 years ago. It's probably an interesting story, but since IEEE754 math is done by hardware now there's not any practical reason for it to be different than it is. – Pointy May 07 '16 at 12:15
  • [Possibly of interest.](https://www.cs.berkeley.edu/~wkahan/ieee754status/754story.html) – Pointy May 07 '16 at 12:16
  • @Pointy, I see, ok, thanks for the reference, I'll check it out. Do you suggest I delete the question? – Max Koretskyi May 07 '16 at 12:19
  • I don't see any reason to delete it. It may get closed eventually. Off-topic questions aren't necessarily bad things, they're just off-topic :) – Pointy May 07 '16 at 12:26
  • [There's also this older question which I think might be exactly the same.](http://stackoverflow.com/questions/19864749/why-do-we-bias-the-exponent-of-a-floating-point-number) The answer on that one is clear as a bell, and it's probably something I knew at some point :) – Pointy May 07 '16 at 12:27
  • @Pointy, thanks, that answer that you linked is probably a good read so maybe mark this one as duplicate? – Max Koretskyi May 07 '16 at 12:34
  • @Pointy, [here is](http://stackoverflow.com/a/2835476/2545680) another good one that answers this question. – Max Koretskyi May 08 '16 at 06:20

0 Answers0