> Is there a reason not to overload at Word32? (I suspect the List.tabulate > bound is wrong in generate-overloads.sml.) You're right. I fixed it. It would be nice to have a regression test or two that checked overloading (and default overloading as well, which I messed up at first).