Code Review in Cabal Project

[Emily Pillmore]

Just to pile on, I read the conversation and thank you Francesco for giving complete and meaningful feedback instead of throwing in the towel on such a large PR. I don't think there's any general means of making sure that contributors have a smooth contribution experience, but in this case, one thing I'd point to is the sheer size of the PR.

As rightfully mentioned by (I believe it was?) Francesco, there are 3 distinct ideas at play, and some were implemented without the contributor knowing the full consequences of making their changes. Had he been asked to split his PR up into digestible, *separate* PRs, I think this may have gone smoother and feedback churn on each component would be kept minimal. When people contribute very large PRs like this, they often drag on and defeat the user's desire to follow them through because they can't get little wins out of the process. There is an upper bound to people's personal tolerance level when it comes to critique. There's a bit of "know thyself" inherent in the process, but we can also ask users at the outset to split Very Big Contributions up as much as possible.

Anyway, sorry I couldn't make the Cabal thing - I have a standup now at exactly that time. If we offset by a week, I'd be able to make all of them.

Cheers,

Emily

On Thu, May 04, 2023 at 1:09 PM, Theophile Hécate Choutri < cabal-devel at haskell.org > wrote:

> 
> Hi Artem, and thank you for the email.
> 
> 
> Yes, thank you very much Francesco, the feedback you left certainly
> avoided a lot of pain, and you contributed to maintain this patch up to
> standards.
> 
> 
> Regarding the "why" and the "how not to reproduce this", interactions that
> leave a sour taste in the mouth are pretty much inevitable in the messy
> business of human interactions. Shit happens, the best we can do is
> reflect on it and do what is in our power to improve what we can.
> 
> But it is true that remote communication, by text, leaves much in terms of
> nuance and interactivity.
> 
> 
> Again, thank you Francesco.
> 
> 
> On Thu, 4 May 2023 at 20:40, Artem Pelenitsyn < a. pelenitsyn@ gmail. com (
> a.pelenitsyn at gmail.com ) > wrote:
> 
> 
>> Dear Cabal team,
>> 
>> I have a couple points on the multiple components PR
>> ( https:/ / github. com/ haskell/ cabal/ pull/ 8726 (
>> https://github.com/haskell/cabal/pull/8726 ) ) drama. I didn't want to
>> spend
>> everyone's time during the meeting, and you can decide if you want to hear
>> more
>> about it now.
>> 
>> The whole discussion on the meeting (with one exception) sounded to me
>> like we're
>> scolding Francesco (@fgaz). I feel bad because of that. Hence my email. By
>> the
>> way, one exception imo is Mikolaj saying that the whole story is his
>> responsibility too. Thank you Mikolaj! I should say that me commenting on
>> red CI
>> probably added to the heat that was already there, and therefore it's also
>> my
>> responsibility. I'm sorry that the contributor got frustrated.
>> 
>> That said, I want to say a big THANK YOU to Francesco. You left
>> high-quality
>> feedback, that myself (and I think others) can learn from. I personally
>> lack the
>> kind of expertise that you applied when doing this review. Please, do this
>> more
>> often! (time permitting, of course)
>> 
>> People keep asking (1) why this happened (contributor throwing hands in
>> the air
>> and seemingly abandoning PR) and how to avoid this in the future (2).
>> 
>> I don't feel qualified to answer (1), and I think no one should try to
>> play
>> contributor's therapist (unless, of course, you have the education and
>> licence
>> for that). I don't believe any professional boundaries were crossed or
>> something
>> unreasonable happened. Not from Cabal team's (including Francesco) side,
>> at
>> least. All the requests were fair and polite requests advocating for
>> improving
>> quality of Cabal's code base -- the very code base we interact a lot with,
>> and
>> quality of which we should try to improve.
>> 
>> My answer to (2) -- you won't be able to avoid it in the future. Unless,
>> of
>> course, you are willing to abolish the code review process altogether. It
>> really
>> depends on a contributor and whether they are able to sustain the scrutiny
>> of
>> the process. And you can't control that. The response shouldn't be to
>> lower the
>> bar for the process imo.
>> 
>> It is my opinion that the biggest improvement to the future of Cabal would
>> be to
>> get more high-quality feedback for contributions, not less. Thank you,
>> Francesco, again!
>> 
>> 
>> --
>> Kind Regards, Artem
>> 
>> 
>> 
>> _______________________________________________
>> cabal-devel mailing list
>> cabal-devel@ haskell. org ( cabal-devel at haskell.org )
>> http:/ / mail. haskell. org/ cgi-bin/ mailman/ listinfo/ cabal-devel (
>> http://mail.haskell.org/cgi-bin/mailman/listinfo/cabal-devel )
>> 
> 
> 
> 
> 
> _______________________________________________
> cabal-devel mailing list
> cabal-devel@ haskell. org ( cabal-devel at haskell.org )
> http:/ / mail. haskell. org/ cgi-bin/ mailman/ listinfo/ cabal-devel (
> http://mail.haskell.org/cgi-bin/mailman/listinfo/cabal-devel )
> 
> 
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/cabal-devel/attachments/20230504/286a7b5d/attachment-0001.html>