Open greole opened 2 weeks ago
I had a look into this, as I was hoping to do a bit of warm-up on it :wink:. I am not sure doing this makes sense, because the boundaryFields
is not a Field
in the same way that the InternalField
is a Field
. I would suggest we actually change the BoundaryFields
name to BoundaryFaceConditions
or FaceBoundaryConditions
or something similar. It does not really represent a field per se; it's more of a boundary condition container that may contain values of size faces.
Yes I strongly agree with renaming boundaryFields
,
BoundaryValueContainer
? Because as you say it is a container that holds all required values to compute the boundary conditions. That would be also my argument against BoundaryFaceConditions
because a. it doensn't know anything about the actual condition and b. maybe we Face could be misleading.
Currently, when calling size() for
GeometricFields
theinternalField
size is returned. While this is often what user intends it is, however, ambiguous. One better solution would be to return:a. a tuple<size_t, size_t> with the internalField and boudaryField size b. a struct GeometricFieldSize which holds both values.
in both cases one has to explicitely select which size is desired.