on Thu Feb 02 2017, Xiaodi Wu <swift-evolution@swift.org> wrote: > On Thu, Feb 2, 2017 at 9:45 AM, Dave Abrahams via swift-evolution < > swift-evolution@swift.org> wrote: > > If indeed the desired semantics for ranges is that they should continue to > lack precise semantics, then an agreement that we are going into this > deliberately and clear documentation to that effect is the next best thing, > I suppose.
Practically speaking, using the type system to separate ranges that are collections from the ones that merely bound some values would not be helpful to anyone, IMO. If that's what you mean by saying ranges have imprecise semantics, then that's what I, at least, desire. A more powerful way to look at it, IMO, is that ranges unambiguously represent one or two bounds on comparable values, and there are operations that combine those bounds with other entities (for loops, collections whose indices match the ranges) in useful ways. -- -Dave _______________________________________________ swift-evolution mailing list swift-evolution@swift.org https://lists.swift.org/mailman/listinfo/swift-evolution
