David W. Hogg

It feels like a constrained-positive solution is possible -- because it is for any finite-parameter model. But I don't know how to do it!

Yes, yes and yes. Happy to help with everything.

In principle, if the dimension of your problem is large (many parameters) you should reduce the size of the step size in the ensemble move. It is the "a" parameter...

If there are other parameters _not_ being sampled by emcee, then I presume you are fixing them within the sampling to the same value for all walkers? Anything else would...

This will not work in general. If you look at the integral you are trying to do, and the integral that is implicitly being done by averaging over samples, you...

This is a long-sought issue but a bit challenging in the current framework.

what do you mean? What's the issue?

I thought you could change branches without any additional pull.

Can we "handle" this exception -- and call out to fitsverify and then say "image fails to fitsverify" and maybe give some further diagnostics?

AVM was never agreed to by the community. And indeed the FITS standard has no display advice. So let's change the title of this issue.