Sean Pinkney

I mean the derivatives are what's hard. At the language level no one needs to know the pivoting. I have an implementation of a cholesky factor of correlation matrices but...

This is also useful for the `hmm_marginal()` function. Where one will have an array of simplexes that need to be converted to a matrix. (I guess `hmm_marginal()`could have an additional...

@WardBrian we have a log_diff_exp. Should we also add a log_subtract_exp?

> CC @spinkney we confirmed that the log(det(jacob) thing works out to the same as the one defined in your algorithm. We tested by doing the `sqrt(log(det(jacobian.T @ jacobian)))` mentioned...

Having the option to do both is ideal. If users also want to inspect the `jacobian` or turn it off in the above function, even having a `target +=` only...

> @spinkney: In your example, why are you incrementing `target` and `jacobian`? Aren't both part of the change-of-variables adjustment? I know part of it's not technically a Jacobian, but think...

> Yes! I have not had time, but what I would really like to do is replace `!is_constant` and `!is_all_constant` with a helper variable template like `is_autodiff_type_v`. Then use `constexpr`...

Are these functions in the new release?

> Other than the first equation, this looks great. Intense, and I have no idea what it's useful for, but the doc looks good. Just wanted to chime in and...

@andrjohns has a really nice copula post at https://users.aalto.fi/~johnsoa2/notebooks/CopulaIntro.html. He also has one about linking base R functions written in C to Stan https://users.aalto.fi/~johnsoa2/notebooks/R_C_Math.html. We should get a discussion on...