Tomás Chor

That's a good point. I missed that option. However, I get an error when that option is activated: ``` TypeError Traceback (most recent call last) ~/ML6mov.py in 41 gif_palette=False, 42...

@jbusecke I gave you my comments on the quick start guide in #271. But that's a very good start! I've read the examples and (at least as far as I...

@jbusecke Responding here to your comment/request on #291. The quick start guide seems like a very good start! It definitely makes understanding how `xgcm` works on a high level much...

> Yeah, this is very confusing, and should definitely be remedied in the future. This will require some rather big changes and a longer deprecation cycle thought (see #195). How...

I would argue that `SpecifiedTimes` shouldn't be actuated in iteration 0 since, if that is desired by the user, it's trivially easy to add zero to the list of times....

I can confirm that I can reproduce the error using the MWE [here](https://github.com/CliMA/Oceananigans.jl/issues/2744#issue-1381826641) with CUDA 3.12. As with #2756, the same snippet appears to run fine with CUDA 3.9 though....

> `w` is not a prognostic field for the hydrostatic free surface model so it cannot be used as a dependency in continuous boundary functions. > > We could think...

> It might make sense to follow @simone-silvestri's suggestion and include all fields in `field_dependencies`. That would open up more options for forcing in general, esp using `auxiliary_fields`. Just to...

Velocities, tracers, and auxiliary fields seems pretty good already. I completely agree with this.

CC @simone-silvestri