[swift-evolution] protocol-oriented integers (take 2)

[Xiaodi Wu]

On Mon, Jan 30, 2017 at 10:08 PM, Brent Royal-Gordon via swift-evolution <
swift-evolution at swift.org> wrote:

>
> > On Jan 30, 2017, at 11:25 AM, Dave Abrahams via swift-evolution <
> swift-evolution at swift.org> wrote:
> >
> >> I mean that `OptionSet.RawValue` currently has to conform to
> >> `BitwiseOperations`,
> >
> > Actually it doesn't.  You just have to implement these yourself in that
> > case:
> >
> >  extension OptionSet where Self.RawValue : BitwiseOperations {
>
> Oh, I didn't realize it was implemented that way (and was going to stay
> that way). Thanks for the correction.
>
> >> but would now need to conform to `BinaryInteger` (or a sub-protocol).
> >
> > Does that limit you in some useful way?
>
> Well, a type like `Data` could be usefully conformed to
> `BitwiseOperations`, which would permit its use as a variable-sized bit
> buffer, but conforming it to `BinaryInteger` would make no sense and might
> cause mis-conformances. (For instance, `BinaryInteger.Type.+` and the `+`
> operator that works on `RangeReplaceableCollection`s like `Data` are
> incompatible). You would instead have to use a big-int type, but it's
> apparently common for those to be implemented in ways that make bitwise
> operations slow.
>

Having implemented a bit vector in Swift (then found it to be slow, then
asked for leadingZeros the last time this proposal came round), I totally
agree that it would be one such type that conforms to the syntax and
semantics of `BitwiseOperations`.

I was initially bummed to see it go as well, but on reflection I think it's
the right decision. The question is, what interesting algorithm could you
actually write that would be generic over BitwiseOperations? I couldn't
think of any. Sure, I could use an Int like a BitVector, but if I wanted to
use the instance only for bitwise operations, I'd just use a word-sized
BitVector. Perhaps you have other use cases?


> However, unless I'm mistaken, I believe a `BitwiseOperations` protocol
> could be extracted from `BinaryInteger` later (right? Resilience permits
> you to add a new super-protocol and move some of the sub-protocol's
> requirements up into it?), so we can pick that up later.
>
> --
> Brent Royal-Gordon
> Architechies
>
> _______________________________________________
> swift-evolution mailing list
> swift-evolution at swift.org
> https://lists.swift.org/mailman/listinfo/swift-evolution
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.swift.org/pipermail/swift-evolution/attachments/20170130/2ca4f98f/attachment.html>