11:39:45easyeHmmm. In the course overhauling the way we store vectors, it seems that our previous implementations of (unsigned-byte 16) and (unsigned-byte 32) were using arrays of int[] and long[] respectively for storage, effectively doubling the amount of memory needed.
11:40:01easyeWonder if there was more subtle reason for this?
11:43:38easyeAh, probably because short,int, and long are all signed quantities, meaning that we would need to coerce unsigned values in and out, so simpler to pick a precision where one can just store the entire value.