The Package Addition Process: a review

This page describes our experience trying to get packages adopted via the Haskell Platform addition process. Particularly, the experience around the 'text' package discussion.

Feedback

Feedback from that effort is gathered below:

  • So far, very few proposals have been made, so the goals to grow haven't been achieved.
  • Lots of review did take place, from many participants. 107 emails were produced in two months, by 29 reviewers. Several detailed reviews took place.

Other comments:

  • I think that the "please speak up only if you dissent" model is quite counterproductive. It gives the strong (but, I hope, incorrect) impression that those who care are all dissenters.
  • Too much bikeshedding took place, relative to the overall goals of the Haskell Platform (to be comprehensive, to grow, and to be high quality).
  • The existing libraries are often in a poorer state than the new libraries, yet the new libraries are possibly subject to blocking objections, despite those flaws existing in the current libraries.
  • It is unclear on what issues people should raise blocking objections, and when conditional acceptance is suitable.
  • The steering committee didn't play an active role in herding discussion, or communicating expectations.

Possible actions

  • Push on, as we get as a community review more things, the cultural norms will become clearer, and the mechanism more efficient.
  • Limit the scope and time for reviewers (e.g. an opt-in list like Haskell Prime)
  • Explicitly state new norms for when something may be significant enough to be a valid reason to block.