Hm. I'm not sure anything's going to change here; strictly maintaining source compatibility would mean never adding new overloads. cc @phausler, @rudkx, @airspeedswift

`init(bytes: Sequence)` and `init(_ elements: Sequence)` are not overloads in that they have two different names. It seems linguistically that fetching the init function should have a more qualified name: e.g. via argument labels as they are part of the signature fo the function.

Consider the following translation to objc:

+ (instancetype)dataWithSequence:(Sequence<UInt8> *)sequence;
// and the other method
+ (instancetype)dataWithBytes:(Sequence<UInt8> *)sequence;
// so fetching the IMP to pass to a mapping function would use two distinct selectors
[Data methodForSelector:@selector(dataWithSequence)];
[Data methodForSelector:@selector(dataWithBytes)];

So in the end it seems like there ought to be a way to specify

init(:)

versus

init(bytes:)

imho.

Also it might be suggested to use

"Hi".data(using:.utf8)

instead anyhow to convert to data from a String.

That's not how Swift works. "Overloads" means the same base name, differentiating first by argument names and then by types. That's important for default arguments.

(You certainly can write string.map(Data.init(_:)), and I kind of thing the current syntax shouldn't be legal myself, but it is. Still, I don't think that should stop us from adding new overloads, especially new initializers!)

Definitely not a huge deal if this isn't source compatible, but I wanted to flag it in case that wasn't desired.