Tomás Chor

> you may want > > using Oceananigans > using Oceananigans.Grids: znode > using Oceananigans.Fields: @compute > > underlying_grid = RectilinearGrid(size = (4, 4, 4), extent = (1, 1, 1))...

> > The docs weren't very clear on what `Integral` expects for `mask`. > > I agree. I think the relevant information is in `ConditionalOperation` docstring > > [https://clima.github.io/OceananigansDocumentation/stable/appendix/library/#Oceananigans.AbstractOperations.ConditionalOperation-Tuple{Oceananigans.Fields.AbstractField}](https://clima.github.io/OceananigansDocumentation/stable/appendix/library/#Oceananigans.AbstractOperations.ConditionalOperation-Tuple%7BOceananigans.Fields.AbstractField%7D) >...

> It's just in the function signature -- you don't have to use it. Sorry about that, try this: > > condition(i, j, k, grid, co) = ifelse(znode(i, j, k,...

> > Yes, I was also thinking about the renaming of condition-mask potential confusion... I do like the kwarg name "condition" but I do think that usually people think of...

I think he is proposing: - `condition` -> `mask` - `mask` -> `mask_value` no?

> You are right; I had misinterpreted. > > Actually `mask -> masked_value`. And I actually like that even more. I think I agree with @navidcy that I like `mask...

At the moment the user has to define `Figure` and a plotting function to create a movie maker, such as: ```julia using CairoMakie fig = Figure(size = (800, 400)) ax1...

Do the changes here have any impact on https://github.com/CliMA/Oceananigans.jl/issues/3485? i.e. will it make it easier or harder to make `Checkpointer` play with time-averaged fields?

> @glwagner, @tomchor so I see in #3485 the discussion/proposal is to have `write_output!(sim::Simulation, ...)` instead of `write_output!(model::AbstractModel, ...)`. > > We can do that here. But then it means...

> This PR includes few useful updates/refactors for Clock and I'm quite tempted to remove the Checkpointer things, merge and continue with Checkpointer in a new PR. What do you...