On Wed, Oct 04, 2017 at 03:20:52PM +0200, Erik Bray wrote:
> On Wed, Oct 4, 2017 at 8:57 AM, Gerard Capes
> <[email protected]> wrote:
> > Dear Software Carpenters,
> >
> > I've had a few PRs open on SC lessons for around 2 weeks. Two of them have
> > not received any response at all. One of them had a brief conversation,
> > ending with me asking for clarification as to what edits would be required
> > to have the PR merged.
> >
> > Now I know the maintainers are volunteers (I am also a lesson maintainer),
> > and have other priorities, but it's frustrating to go to the bother of
> > submitting a PR only to get nothing of use back.
> >
> > So I'm asking two things:
> >
> > Should Software Carpentry have a response time, where a maintainer should
> > assign the PR within 2 days, or at least comment to indicate their expected
> > response time? This is the approach I have been taking, but I must concede
> > that I get far fewer PRs on the lesson I maintain.
> > What can PR submitters do in my situation?
> 
> I don't have too much to add on what April already wrote, but this is
> a perennial problem on volunteer-based projects, or even on project
> that have paid contributors (a few projects I work on have paid
> contributors, but who are still stretched too thin to always be timely
> in reviewing PRs).  As such, I've had tasks that should have at most
> taken a few weeks take *months* just due to the lack of time on the
> part of people qualified to review my work.
> 
> I have very mixed feelings about all this.  I'm 100% in favor of a
> general requirement that all changes go through a review process.  But
> perhaps there should be a time limit after which at least *trusted*
> contributors (by some measure) should be able to merge changes that at
> least haven't already received negative response.  This can be
> problematic too though.  Because sometimes no response just means
> literally no one else has seen it.  It might be nice if GitHub had a
> way to loudly call attention to PRs that have yet to even be given a
> +1 or -1 by someone...

Hi Erik & others,

tl;dr? (a) add maintainers liberally, (b) forgo review, (c) trust people.

So, we ran into this problem on the khmer project a while ago - it's
unreasonable to expect grad students and postdocs to pay regular attention to
code review.

Three approaches we've tried/are trying:

---

first, add maintainers liberally, e.g. use something like the Pull Request
Hack to give people responsibility when they demonstrate capability:

https://felixge.de/2013/03/11/the-pull-request-hack.html

---

second, use something like Lazy Consensus to forgo review after some period
of time,

http://nowviskie.org/2012/lazy-consensus/

by saying "well, if no one cares enough to comment after a week, go for it!"

---

third, pay someone to review pull requests and fix bugs.

We pay Tim Head (@betatim) to "watch" our github repos.  This works quite
well and (on the scale of things) is quite inexpensive, but is still some
$$.

---

I don't think forgoing review is a good idea, especially with the current
way lessons are managed (merges are essentially "live" immediately, with
no staging area/review before release).  So #2 is not a good idea for SC/DC
IMO.

From my perspective, I'm not sure there's a hurry here, other than (apologies)
the impatience of people who just want to get on with it.  Understandable,
but not something that can be fixed easily!

best,
--titus
-- 
C. Titus Brown, [email protected]
_______________________________________________
Discuss mailing list
[email protected]
http://lists.software-carpentry.org/listinfo/discuss

Reply via email to