Zachary P Christensen

#307 tests all the examples in this issue. I'm assuming there are still times when we don't detect dense dimensions just because we can't know in some situations without knowing...

The only trait in base we have is `RangeStepStyle` which isn't extensively used and has variable interpretations (see https://github.com/JuliaLang/julia/pull/35279). It may also be useful to have traits that are based...

@chriselrod I like this approach a lot. It would also be worth having ```julia known_step(::Any) = nothing known_step(::Type{

After a month of dedicated [bike-shedding on discourse](https://discourse.julialang.org/t/commit-to-a-common-syntax-for-accessing-additional-index-mapping-information-on-axes/84988), I think this is ready for review. It would be good to have some relevant package maintainers okay the name of the...

@mcabbott, would you mind providing some input if you have time?

We have `parent_type` now. You can do `parent_type(T)

I'm assuming these issues are arising from manually writing out types like `StaticArray{Tuple{3,2}}`. I'd suggest instead dispatching on the return of `ArrayInterface.size(x::StaticArray)`, which should be just as fast as dispatching...

size was never intended to be called on the type. We should probably document this

The first example here is fixed. Did we want an error on the second one? This is what we get now: ```julia julia> ArrayInterface.strides(vec(A')) ERROR: ArgumentError: Input is not strided....

I'll try to respond in more detail later today when I have time, but for now I'd say the biggest issue is that we are missing basic tools for memory...