Wei Zhang

Apparently I do not know what's happening here. But when I made AD for those distributions, I did check that everything compiles ok. Maybe @perrydv @dochvam you have better ideas...

> I've got it working up through executing the tests, but test-AD.R is failing. Hey Ben @dochvam, I updated the tests for N-mixture models which passed on my machine with...

Totally agreed on the AD and non-AD opinion, @perrydv. And yes a good idea to keep the non-AD functions and add AD ones as new.

@paul-vdb Thanks for the example. I can run the code fine on my machine - the last line does not lead to crashes. The results do not look good though....

@paul-vdb @perrydv If there is only one param/random effects node, we expected one to use `getNodeNameSingle()`. But now I see the confusing point is that the user may not know...

Thanks @perrydv. I found that declaring N as double in the distribution then the first example worked. The issue in the more complicated scenario was that N was also used...

@perrydv I think I figured out what was happening in the N-mixture distributions where I thought I should declare `Nmin` and `Nmax` as integers instead of doubles. The compilation problem...

`buildQuad`? AGHQ is of a particular type and the more generic name `Quad` for quadrature will allow us to build in other types if of interest?

The changes look good to me. :) I fixed a wee bug in setupMargNodes in `nimbleQuad`. Not sure when Laplace will be reallocated to `nimbleQuad`, so I just fixed it...

All looks good to me @paciorek. I just edited the `last.best` description a bit - to me, it means that the optimised random effects values that correspond to the currently...