Apologies, I was coming at this from the widely used currency type side of things not the code size side of things. The Foundation import wouldn’t be too hard to fix with a trait. Many currency types (such as paths) benefit from having some default Foundation interoperability to make them easy to work with out of the box imo, so an opt-out Foundation trait seems reasonable for situations where foundation is undesirable (a relatively small/advanced subset of use-cases in today’s Swift ecosystem).