My point was that NAG does. I think there’s a flag to switch from 1, 2, 3 to 4, 8, 16 —my sycophant-LLM-of-the-day seems to confirm/hallucinate it, hehehe.
After trying PDTs many times, but ultimately hitting a wall and doing something else, I’m now convinced that PDTs are conceptually broken… it only took us 20+ years to find that out.
That’s useful for generated code, and it also unifies the syntax into the tyoe-guard of the select-type-construct.