I agree that these make sense, but considering how late this SRFI is, I want to hold them for another SRFI.  I'd put the BE/LE into the procedure name, though.  The (rnrs bytevectors) library can do all the grunt work, as for this SRFI.

On Mon, Aug 5, 2019 at 10:40 AM Lassi Kortela <xxxxxx@lassi.io> wrote:
Procedures like this are conspicuously missing:

(u8vector->s32vector-copy 'little-endian foo 0 16))

This would treat bytes [0,16[ of the u8vector foo as four signed 32-bit
values in little-endian byte order. Do you think this would be too fancy
/ not useful enough?