Open Clonkk opened 3 years ago
For N > 1
I think the only reasonable option is to go with Tensor
s. Otherwise we have to create separate overloads for each N
if we go with tuples. Arraymancer basically does that up to 6 dimensions but we could have to handle many more dimensions here.
... And how should we handle them ?
Should it based on seq of tuples ? On Tensor ?