[ghc-steering-committee] GHC 2020

[Joachim Breitner]

Hi,

Am Montag, den 19.10.2020, 14:07 +0000 schrieb Richard Eisenberg:
> * This counter-counter-proposal is meant to combine what I see are
> the best qualities of both proposals: the structure in Alejandro's
> (Joachim's suggests just debating about extensions over email, which
> causes me to shudder in horror, even if the discussion is just among
> committee members), with the committee-seeding from Joachim's.

My proposal is mis-phrased: I suggest _not_ debating. I want GHC2021 to
contain only uncontroversial extensions. If debating flares up about a
specific extension, that’s already a clear sign that the extension
should not make it _this round_, which settles that very discussion.


If at any point a committee member feels like strongly arguing in favor
of an extension that does not already have a majority with more
emphasis than just “here are some good points you might have missed”,
then such an extension is not uncontroversial, and would simply be
voted out. I expect us from refraining from strongly and verbosely
arguing for an individual extension, and thus no discussion ensues.


If at any point a committee member feels like strongly arguing against 
an extension that does already have a majority, then I expect that at
least three other members will take that as “clearly not
uncontroversial”, remove it from their vote, and make any further
discussion not needed.



I think it is wrong to think the decision “Should Foo be part of
GHC20201” needs the same level of rigor as “should Foo exist”.

In the latter case, if we say no, somebody is unhappy because they
can’t use a feature they care about. So we really need the detailed
discussion, and refinement etc. So yay for PRs.

In the former case, if we say no, it’s no big deal, people will have to
continue saying {-# LANGUAGE Foo #-} for a few more years. This does
not hurt anyone!


And therefore, I strongly advise against any heavy per-extension
discussion process.


> We'll all edit the meta-proposal as need-be, with the expectation
> that we friends do not clobber each other's work. Seems simpler than
> having the conversation spread among multiple PRs.

That’s a good point; we can have options in one PR and still do ranked
voting.

Cheers,
Joachim

-- 
Joachim Breitner
  mail at joachim-breitner.de
  http://www.joachim-breitner.de/