[Returned for revision] SE-0270 Add Collection Operations on Noncontiguous Elements

The review of SE-0270: Add Collection Operations on Noncontiguous Elements has concluded.

The review yielded a mixed signal:

  • There was significant support for providing a set data structure of arbitrary (e.g. non-Int) indices.
  • There was significant confusion (including on my part!) about the intended meaning of the proposed set type, which was documented as having two different element types.
  • There was almost no feedback about the stated topic of the review, “Operations on Noncontiguous Elements” which—in fairness—did not constitute the bulk of the proposal.
  • There was lots of disagreement about naming choices, driven at least in part by the unclear answer to the “what is this thing?” question.

In the end I spent several days just analyzing the design space and writing code, and got back to @nnnnnnnn with my findings, which were also mixed. Out of that we see a promising direction that we'd like to take back to the pitch phase. We also think the proposal should be broken into smaller pieces so that each can receive the attention it deserves. The core team agreed with our conclusions, so the proposal is returned for revision.

Thanks to all for your engagement and commitment to the review process. Every review result, no matter the outcome, is forward progress on the evolution of Swift!

Dave Abrahams
Review Manager