I have a template, will add it.

We don't need much, just enough for a passing observer (or a "Research Note"). I can add this when I add my notes in #2

I learned how to do this at AHW2016, where Dan FM made the following Travis file:

https://github.com/davidwhogg/mcmc/blob/master/.travis.yml

Basically we have do divide out the prior we applied (given alternative model A) and retrospectively apply the prior we want, P(z|h,H) - but it seems this could end up meaning re-weighting by 1/L... I have notes to check and type up.

@drphilmarshall and I were talking about the paper today and identified a potential problem with Eq. 2. An integral over d_i makes this an integral over one galaxy's photometry. The sum of Eq. 3 would then be over all values within D for the photometry of a single galaxy, not over the observed photometry of every galaxy in the survey.

We also think that z_i should just be z in Eq. 2 so the left and right sides at least have the same variables on each side of the conditional. If that was the intention, Eq. 2 may also run afoul of Eq. 15 from here. @davidwhogg might be interested in weighing in here.

Hi @aimalz ! Here's the project I was telling you about last week at the LSST DESC Hack Week. Interested to hear your thoughts on our notes - marginalizing over datasets was a new concept to me, and seems like an interesting thing to think about.