On 2026-01-11 22:02, Jacob S. Gordon wrote:
> On 2026-01-10 03:57, stardiviner wrote:
>> Why not support float number in this statistics cookie? It's more
>> meaningful and human readable.
>
> That would help distinguish small percents, but it’s a bit orthogonal
> IMO since there’s still the question of how it’s rounded. If the
> percent cookie precision is p then with N = 10^{p+2} + 1 the fraction
> 1/N could still round to 0, e.g. with a ‘%.pf’ format specifier:
Actually, the low-end works, my mistake:
|---+-------+----------+------|
| p | N | 100/N | %.pf |
|---+-------+----------+------|
| 0 | 101 | 0.990099 | 1 |
| 1 | 1001 | 0.099900 | 0.1 |
| 2 | 10001 | 0.009999 | 0.01 |
|---+-------+----------+------|
#+TBLFM: $1=@#-2::$2=10^($1+2)+1::$3=100/$2;%.6f::$4='(format (format "%%.%df"
$1) (/ 100.0 $2));N
it’s the other end that gets rounded ‘incorrectly’:
|---+-------+------------+--------|
| p | N | 100(N-1)/N | %.pf |
|---+-------+------------+--------|
| 0 | 201 | 99.502488 | 100 |
| 1 | 2001 | 99.950025 | 100.0 |
| 2 | 20001 | 99.995000 | 100.00 |
|---+-------+------------+--------|
#+TBLFM: $1=@#-2::$2=2*10^($1+2)+1::$3=100($2-1)/$2;%.6f::$4='(format (format
"%%.%df" $1) (/ (* 100.0 (1- $2)) $2));N
Either way, my point was just that the precision and rounding
behaviour around 0/100% can be independent.
Best,
--
Jacob S. Gordon
[email protected]
Please don’t send me HTML emails or MS Office/Apple iWork documents.
https://useplaintext.email/#etiquette
https://www.fsf.org/campaigns/opendocument
