bug?

[Qyokizzzz]

[bartaz]

Thanks for taking time to report this.

To be honest I created it almost 10 years ago and I really don't remember the exact details. Certainly this seems like a bug (exponent is 0, but equations suggest it's 1), but on the other hand it is treated as 0 inside the sum in brackets.

It would be really hard for me now to try to understand it again and fix that. If I get a chance I'll try to have a look into it (it may be related to issue mentioned in #2). But I can't really promise anything.

Anyway, thanks again for using my tool and reporting an issue! I hope it was helpful :)

[bartaz]

@Qyokizzzz I found the part of the Wiki article that explains this: https://en.wikipedia.org/wiki/Double-precision_floating-point_format#Exponent_encoding. So, the behaviour is explicit and correct, because when exponent is 0 it is treated as special case and the standard equation doesn't apply.

But indeed this is confusing that the last line of equation says 2^e-1023 while it's really 2^1-1023.

[Qyokizzzz]

@Qyokizzzz I found the part of the Wiki article that explains this: https://en.wikipedia.org/wiki/Double-precision_floating-point_format#Exponent_encoding. So, the behaviour is explicit and correct, because when exponent is 0 it is treated as special case and the standard equation doesn't apply.

But indeed this is confusing that the last line of equation says 2^e-1023 while it's really 2^1-1023.

It should be a typo.

Wikipedia says "00000000000 is used to represent a signed zero (if F = 0) and subnormals (if F ≠ 0)".

That means it represents all underflowed numbers when the exponent is 0, so 00000000001 is the special value to cause the number before fraction is 0.