[question]: Some functions can panic

[nhatcher]

I see some functions can panic. For instance:

https://github.com/statrs-dev/statrs/blob/master/src/distribution/beta.rs#L191

Is that something that might change in the future or that is a design decision?

This makes things a bit difficult for us. See
ironcalc/IronCalc#654

Apologies if this is not the appropriate channel and since I am here, congratulations on a great crate!

[YeungOnion]

This is a good channel for this feedback - I've just been fairly absent as of late.

Previously, the DX requiring a newtype as an argument (just an f64 that implements num::Bounded) to these functions didn't seem good and after it would have led to breaking change for the inverse_cdf API which seems to be one of the most used parts of the library.

Since you wrap this functionality for multiple distributions, I'd love to get your take on how you'd like to consume the API. Are you binding each callable or are you binding to the ContinuousCDF trait?

[nhatcher]

This is a good channel for this feedback - I've just been fairly absent as of late.
No problem!

I'd love to get your take on how you'd like to consume the API

I don't really have a preference, as long as it cannot panic.

Since you wrap this functionality for multiple distributions, I'd love to get your take on how you'd like to consume the API. Are you binding each callable or are you binding to the ContinuousCDF trait?

This is an example of how we use it presently:
https://github.com/ironcalc/IronCalc/blob/main/base/src/functions/statistical/chisq.rs