Sam Schlinkert

For Shannon line max length: Using 6.1_M_, then rounding **down** to nearest integer (can't have fractions of words on the list), I get | shortest word length | max list...

> you'll have to decide whether you consider 2.6 or log26.1 (or even something else, like 2.63, or 11.82/4.5) to be the canonical representation of Shannon's limit. Glad we finally...

> to me, this suggests that Shannon's notion of "entropy per letter" is not suited to characterizing the vulnerability of word lists to letter-based brute-force attacks. I think you're asking...

I think I follow you, but I want to learn/know more about _G_, with an eye toward hopefully (a) improving its name or (b) mathematically converting into something more intuitive...

> What I'm less clear on right now, is how to interpret all of this when there is a non-uniform frequency distribution in the corpus. I've seen some references to...

Oh, this is helpful. I think I've gotten some of the benefits with this commit: https://github.com/sts10/tidy/commit/157141ffdc5c66c4da7fc69bcda8eb7a41c4e7d1 . Basically, if user does not specify a `--locale`, the program will attempt to...

> "desirability" is going to make everything more complicated. I _think_ I can keep it relatively uncomplicated by using the order of the inputted word list as a proxy for...

There's so many things to check! I'm trying to think bigger here: If we keep adding attributes, I think it might be best to separate these functions into two different...

Haven't seen this issue since I opened this issue. Will leave it open if other users or future me runs into it again.

Recently, [it seems that an Apple employee requested that the developers of popular text editor nano re-license the editor under a permissive license rather than GPL3](https://lists.gnu.org/archive/html/nano-devel/2023-04/msg00003.html). So far, the developers...