comparison with broom.mixed

[bbolker]

So far I have been continuing to actively maintain broom.mixed. However, I wouldn't object to putting it into maintenance mode (probably not removing it from CRAN for quite a while) if it seems to be superseded by parameters.

Thoughts? Pros and cons?

[bwiernik]

I think parameters provides generally equivalent and broader support compared to broom.mixed, so I think maintenance mode (like broom) makes sense

[strengejacke]

Some pros/cons:

From a developer perspective, parameters has a bit of "overload", but there are arguments to avoid the time-consuming formatting of the output data frame, thus, computation time for tidying should not be crucial.

The major difference is that in easystats, the tidy-stuff is in parameters, the "augment/glance" (?) stuff is in performance, while you have both functions together in broom / broom.mixed.

broom is probably still used more widely by users directly, while meanwhile some wide-spread packages like modelsummary or gtsummary rely on parameters and performance in first instance (and fall back to broom only if the other fail).

[strengejacke]

From a user-perspective, parameters might be more convenient to the additional features, as Brenton mentioned, like standardizing, different ci-methods etc., and nice printing.

[strengejacke]

Are there any features in broom / broom.mixed that are not yet implemented in parameters?