[J3] Question on TINY example
Steven G. Kargl
kargl at uw.edu
Thu Jun 23 05:55:19 UTC 2022
On Thu, Jun 23, 2022 at 07:15:15AM +0900, Malcolm Cohen via J3 wrote:
>
> The description is correct. It follows from the model in 16.4. That model has 1/radix <= mantissa <1. So the minimum number is radix**(emin-1) QED. (This is not how IEEE-754 describe themselves, they have 1<=mantissa<radix.)
>
>
> To all: these descriptions were written back in Fortran 90 days by numerical experts. That does not mean that they always got it right, but you should think twice, thrice, and then again before assuming that they are wrong.
>
Given that at least 3 members of J3 mailing list confused the model
in 16.4 with IEEE-754 binary24, 22-007.pdf spends 37 pages in Section
17 discussing IEEE floating point, and the numerical experts from
1980s may no longer be available to provide insight, it would
seem prudent to include a statement in 16.4 NOTE 1 to clearly
identify what the model in 16.4 is.
https://dl.acm.org/doi/pdf/10.1145/355972.355975
--
Steve
More information about the J3
mailing list