(1) [...] but that a multi-argument `.at(x,y,z)` would be completely novel in the STL and therefore very clearly "design work" not to be snuck in by LWG; it needs a paper. Such a paper seems non-controversial to me, but who knows.

This is a great point, thanks! I agree that variadic templates `at(...)` are *new* to the standard, and when I try to spell out the "*Effects*: Equivalent to:" wording that fits it, it's not as straightforward as it seems, which is indeed something worth considering.

So I think `mdspan::at` deserves its own paper, perhaps discussing the rationale for further providing `front()` and `back()`? Because in my opinion there is also value in these two, the former eliminates the need for spelling `ms[0,0,...]`, and the latter simplifies `ms[ms.extent(0)-1,ms.extents(1)-1,...]`, which may be convenient for users.