[Haskell-community] Civility notes (was "Traversable instances for (, , ) a b")

lennart spitzner lsp at informatik.uni-kiel.de
Fri Apr 28 09:49:54 UTC 2017


A couple weeks earlier there was a discussion on tuple instances on this list
that got somewhat out of hand, leading to a meta-discussion on civility.
There was the suggestion to create and endorse a CoC for this community.

Now both topics have not received much further contribution, an indication that
not much more can be gained from these discussions. Yet I have a bad
feeling about leaving them in such a manner, because: There is no real
conclusion, there is no agreement, and I do not see much advancement of how
we, as a community, cope with negative situations. And while I can understand
that there is little incentive/motivation to continue due to negative
emotions involved, I also fear that ending discussions on such negative
emotions will discourage contributions in general not only now, but in the
future as well.

So I will dare to continue, ask a couple of questions, and make some
suggestions:

1. At which point of the particular tuple instance discussion would it have
   helped to have some CoC, and in what way? Is the hope that the participants
   had considered this CoC and not said something in the way that they did?
   Or would it have allowed us to quickly point out the CoC at some specific
   point in response to some mail? Or something else?

   I _can_ see a couple of instances where a CoC could have been pointed out,
   but these don't convince me, because
   a) in those cases giving clear, respectful negative feedback (for example
      regarding "joking") (would/should) have worked just as well if not better
      and
   b) because simply pointing out the CoC during a discussion is rather
      non-constructive because it is a vague form of criticism and the
      receiving party will most likely consider it inappropriate, and so it has
      the opposite effect.

2. on a related note, I have a hard time pinpointing the moment in the
   discussion where things transitioned from cool to flaming. I'd perhaps name
   as important factors the useless rhetoric (go and ask those mathematicians)
   and the case of hiding behind "it was a dumb joke" followed by what in my
   eyes reads like a dishonest apology. But I am not certain and perhaps
   unfair.

   My subjective estimation is that discussing this a bit further is more
   constructive than working on a CoC. What parts of the discussion were
   unfortunate, exactly, and why? The general opinion here seems to be to
   ask for civility without naming names. I disagree: I have little hope that
   giving the vague feedback to all participants that some parts of the
   discussion were non-constructive/disrespectful will improve things in the
   future.

   As an example, we might take the following advice from this:
   "Humour is important and generally welcome, but it is necessary to be
   especially careful to make it clear when exactly we talk in jest, and to
   not let slip phrases that can easily interpreted as offensive if not
   interpreted as a joke. We will not accept retroactively hiding behind
   'it was a joke'."

   (perhaps some people think such a statement belonged in a CoC, but then
   this is a different/more specific kind of advice than what I can see in
   existing/proposed CoCs.)

3. And back to first discussion: I refuse to vote -1 or +1, because the topic
   is more nuanced than that. Instead, I vote for the following:
   "Additional tuple instances shall be added after such a point in time where
   either the methods have been renamed as to avoid confusion, or after the
   generic versions are no longer exposed in the default Prelude.
   (and whether this point will come is intentionally left open.)"

4. And reflecting on the previous point, I encourage all participants to try to
   not make pure -1/+1 votes, but to include conditions under which they may
   switch, especially for controversial subjects. I have hopes that this will
   help finding a majority-backed compromise.

5. It would help to have the discussion and the arguments made by both sides
   archived somewhere other than on the mailing list. In one of the last
   mails I wrote to this list I implicitly complained about the
   signal-to-noise, and to be clear, I don't mean that any messages consist
   of noise. But it can easily take a couple of mails back-and-forth to get
   some point across, and these threads can grow to over a hundred mails
   quickly.
   I realize that the main issue here of course is the amount of work it would
   mean to somewhat objectively summarize an (often heated) debate. But then
   the alternative is the reiteration of the same topics in an almost
   predicable frequency.
   Thoughts?

(Sorry, Tony, for somewhat singling out the "joking" as the negative example.
This might be unfair. You have a valid point, but conveyed it rather poorly
especially to the end of the discussion.)

-- lennart


More information about the Haskell-community mailing list