You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
so there's some redundancy (perhaps we should export one set to reduce namespace pollution) but a minor practical issue is that the Flux type conflicts with the popular Flux.jl package (I guess I'm the only one using them together right now but there may be more in the future?). There's also some inconsistency in exporting NormalFlow but not NormalFlowBoundaryCondition.
In deciding on what to export for the user interface (see #1132) I'm wondering what do people think about only exporting the long-name version, e.g. FluxBoundaryCondition instead of Flux.
I think this will have a few benefits:
Lower probability for conflicts. Flux is one example, but Value and Gradient are pretty generic terms so I wouldn't be surprised if they conflict with exports from other packages future users may want to work with.
Scripts might read more intuitively, e.g. because you say "a flux boundary condition" and not "a boundary condition of type flux".
If we decide to export the complete set of boundary conditions we could do it by exporting PeriodicBoundaryCondition without having to worry about conflicting with the Periodic topology.
Currently the
Oceananigans
module exportsso there's some redundancy (perhaps we should export one set to reduce namespace pollution) but a minor practical issue is that the
Flux
type conflicts with the popular Flux.jl package (I guess I'm the only one using them together right now but there may be more in the future?). There's also some inconsistency in exportingNormalFlow
but notNormalFlowBoundaryCondition
.In deciding on what to export for the user interface (see #1132) I'm wondering what do people think about only exporting the long-name version, e.g.
FluxBoundaryCondition
instead ofFlux
.I think this will have a few benefits:
Flux
is one example, butValue
andGradient
are pretty generic terms so I wouldn't be surprised if they conflict with exports from other packages future users may want to work with.PeriodicBoundaryCondition
without having to worry about conflicting with thePeriodic
topology.X-Ref: #1132
The text was updated successfully, but these errors were encountered: