Replies: 3 comments 1 reply
-
To kick things off I'll make the following suggestion. We could add one of these short comments:
We would then need to add a more detailed explanation someplace. Perhaps in the |
Beta Was this translation helpful? Give feedback.
-
I very much agree that adding this documentation will be useful, thanks for suggesting in @mlstowell !
Both of these sound good to me. I like the idea of adding |
Beta Was this translation helpful? Give feedback.
-
Documenting attribute array usage systematically is definitely a good thing. Note that we already have these static methods: Lines 796 to 803 in 5bf8d8a int base = 0 .
Regarding the terminology 1 and 2: some places in the code may use the term "list of attributes" meaning the same as "set of attributes". We can probably use these two interchangeably since the ordering implied by "list" is probably not important in most cases. The "marker array" is already used in some places, I think, at least in the context of dofs. For clarity, we may need to explicitly refer to 2 as "attribute marker array" to distiinguish it from "dofs marker array", to avoid potential confusion. Another context where we use such arrays is for mesh elements ( |
Beta Was this translation helpful? Give feedback.
-
MFEM has two standard means of passing element/boundary attribute sets through function calls.
Mesh::attributes.Max()
orMesh::bdr_attributes.Max()
where non-zero entries identify relevant attributes.It is not always obvious which of these schemes is required by a particular function. I would like to add a concise statement to each relevant function which would make this clear. Before making any changes to the documentation I'd like to discuss the best approach here and find a consensus.
Beta Was this translation helpful? Give feedback.
All reactions