Jutho

No it's really only `StridedArray`. Permutations are not stored in compile time information (at least not in the latest master version), since they are dealt with using stride calculations at...

> Did you ever consider contributing your faster code to Base, and putting it in permutedims!(::StridedArray, perm) in base/multidimensional.jl? The problem is that it's actually quite a bit of code,...

@cortner, `similar_from_indices` and `contract!` are two of the functions one needs to implement to make `TensorOperations` work on custom array types, together with `add!` and maybe `trace!`.

Sure, no problem to leave this open. I think it's certainly feasible, I am not sure whether you will be able to get optimal code for `StaticArrays` (i.e. completely/partially unrolled...

Yes, I changed how this works in the latest release but did not update the docs accordingly. Provided that your custom tensor type listens to the call `Base.similar(some_tensor_object, some_eltype, some_structure_specification)`,...

That's something I specifically added for my use case in TensorKit.jl, where tensors have two sets of indices, and you can permute arbitrary among and in between them. For most...

This used to be on the package README but I guess it moved to the docs when I created a more extensive documentation (which has grown a bit out of...

What do you mean by package documentation? There are doc strings for the most common functions and macros. The module itself does not have a docstring. I can add one...

You could leave this open; I haven't had time for it but I am not opposed to doing this, so leaving it open is a good reminder.

Do you have a good tensor notation convention for denoting this operation? I/You can of course easily write a function for that and to the necessary permutations and reshapes, but...