Add DN as new unit - #11591

(Taking this out of the code review since perhaps a discussion that should not be just "resolved away")

On equivalencies and/or whether the units should be the same, I'd be more inclined to go the equivalency route, since at least to me count suggests something with Poisson statistics, while that would not be true for ADU (would need to correct for gain). In that respect, perhaps DN should just be its own unit as well?

As for equivalencies, right now it is not easy for units to have built-in equivalencies (though it is possible; it is used for the magnitude units, and have been meaning to extend that, if only for Celsius and Fahrenheit). It would of course be quite easy to add an equivalency for all of them, say, u.equivalent_counts, though it is not too difficult as is:

c = 10*u.ct
print(c.to(u.adu, equivalencies=[(u.count, u.adu)]))
# 10.0 adu
u.add_enabled_equivalencies([(u.adu, u.ct)])
print(c.to(u.adu))
# 10.0 adu

p.s. Just to be sure that we don't let the discussion prevent a simple addition: for now, we could just stick with a simple addition of DN as its own unit (without prefixes!).

for now, we could just stick with a simple addition of DN as its own unit (without prefixes!).

That seems like a good solution to me. I'll update this PR. Thanks for the discussion and example u.add_enabled_equivalencies().

Would both "dn" and "DN" be useful mnemonics for what should be the base u.dn unit?

FWIW I don't think DN is the same as count. Even among HST instruments, that is not always true ("count" could be e/s!).

Yes to no prefix to DN. Removing prefix for count should be a separate PR.

I think having both dn and DN as aliases makes sense.

While, u.N is taken, luckily it doesn't take D as prefix. 😌

Note to reviewer: If you don't think we need What's New entry for this, please remove the label, thanks!

Super, thanks! Will now merge.