Efficiency concerns #2
Description
Varjag mentioned some efficiency concerns on reddit. While this library hits the "known, reliable precision and rounding behavior in a known range and domain" use case for fixed point, it misses on compact representation and efficiency gains that can come from that and SIMD operations on arrays of these types.
While the compact representation is partially approached by declaring the underlying type as a ranged integer type when low and high limits are provided, any compactness is lost due to the data being stored in a CLOS instance.
SIMD operations are infeasible without a compact array of the underlying type. To approach this, it may be worth it to define a fixed array type.
Also, consider looking at improving the efficiency of the generic operations with inlined generic functions.