review
/
2022-1
2
0
Fork 0
Code Issues 3 Pull Requests Projects Releases Wiki Activity
2022-1/document.md

2.2 KiB
Raw Blame History

SR 2022-1: Scheme Review

Author

Lassi Kortela

Status

Living document

Abstract

This document defines the current Scheme Review process. It is expected to change indefinitely as the process changes.

Rationale

Scheme Review was started as a response to the success and shortcomings of the SRFI (Scheme Requests for Implementation) process. SRFI is over 20 years old so a lot of evidence has been accumulated.

What has worked well with SRFI:

  • Everything is public. Transparency builds trust.

  • Drafts are announced. Keeps regulars active.

  • Separation between authors and reviewers. Provides a good balance between authorial control and accountability to the community.

What has caused trouble:

  • Tight focus on "Requests for Implementation" coupled with the fact that anyone can send new proposals has resulted in a flow of experimental work that has not had time to mature.

  • Many interesting proposals are either fit into the process like square pegs into a round hole (with sub-par results), or left outside the process (devoid of the attention they deserve).

  • Tight deadlines commonly fly by. (90 days is not a lot.)

Specification

[Meant to reflect the current state of things. Not meant to be finalized.]

Each proposal has one or more authors.

The proposal starts when the authors submit a first draft.

The proposal is submitted as a git repo. (Currently hosted at gitea.scheme.org/review.)

The authors are free to send any number of additional drafts.

The authors are free to add more co-authors for later drafts.

Proposal ID

Each proposal is uniquely identified by an ID of the form YYYY-N where:

  • YYYY is the four-digit year when the first draft was received, and

  • N is a running number covering all proposals started in the same year.

The running number starts from 1 since there are some problems with SRFI numbers starting from zero. (Hard to remember that SRFI 0 exits; programs cannot use 0 to mean "none".)

The running numbers are monotonically increasing within a given year, but it's permissible for there to be gaps in the numbering.

It is expected that many proposals take more than a year to finish. No matter how many years it takes, the proposal is still identified by the year when the first draft was sent.