Open mateuszbaran opened 2 months ago
Sounds like a very good idea WE should – towards a 1.0 – focus on reducing ambigutiies to some extend I think.
2 Levels is great, then we could also do something like
injectivity_radius(M; retraction_method=...)
and injectivity_radius(M,p; retraction_method=...)
so that the higher level gets nicer to use. Wr just have to find a good name for the llower level then and could do
new_name(M, nothing, method)
for the first and new_name(M, p, method)
, to still have the global case well-distinguished. Or even have 2 names here.
The current interface makes it easy to cause ambiguities with the two-argument variant, see https://github.com/JuliaManifolds/Manifolds.jl/pull/722 . Maybe the function could be restructured to make it less likely. For example by splitting it into a two-level scheme maybe?