I am the author of this presentation. Feel free to ask me questions and send me corrections to the pdf and youtube video.

Here is are the slides: http://uniforumchicago.org/slides/bash_2025-03-25.pdf

Here is the recording of the video: https://www.youtube.com/watch?v=DvDu8_A2uhs

Here is the stringent.sh library that is shown in the presentation: https://github.com/pottmi/stringent.sh

Great presentation. I've been writing bash for waaaay too long, and, while I nodded and said yup, yup, a few times, I also learned some new tricks, so thank you!

The social graces having been observed, if you were to update this presentation to reflect modern bash, and especially reference variables, what would be your top 3 additions?

(I am a big fan of reference variables. They make it possible to, among other things, do some funky things that would otherwise require liberal use of eval, which I eschew. Figuring out how to ensure that the target of a reference variable was in fact an associative array while running under `set -u` was a trick of which I am proud. Note: what I did works and isn't totally ugly, there may be a better way, YMMV, IANAL, etc.)

Second Q, if you will permit: Given the memory leaks in associative array management that have been corrected over the last several years, what are your thoughts on a) whether or not bash should have ever gone that route, and, potentially more controversially, b) the current state of bash maintenance?

(I'm not going to ask about embedding multiple levels of ${...#...} or ${...#...}, because even if it worked, readability would tend to 0 very, very quickly. I'd rather be repetitive and readable and maintainable. Well. Most of the time.)

Is there an https link to this pdf? So rare to see http nowadays.

Thanks for carrying a 20 year torch on bash maintainability!