[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [MirageOS-devel] [PATCH v2 5/6] Add guide on Communication Best Practice

On 27/09/2019, 10:14, "Jan Beulich" <jbeulich@xxxxxxxx> wrote:

    On 26.09.2019 21:39, Lars Kurth wrote:
    > +### Verbose vs. terse
    > +Due to the time it takes to review and compose code reviewer, reviewers 
often adopt a
    > +terse style. It is not unusual to see review comments such as
    > +> typo
    > +> s/resions/regions/
    > +> coding style
    > +> coding style: brackets not needed
    > +etc.
    > +
    > +Terse code review style has its place and can be productive for both the 
reviewer and
    > +the author. However, overuse can come across as unfriendly, lacking 
empathy and
    > +can thus create a negative impression with the author of a patch. This 
is in particular
    > +true, when you do not know the author or the author is a newcomer. Terse
    > +communication styles can also be perceived as rude in some cultures.
    And another remark here: Not being terse in situations like the ones
    enumerated as examples above is a double waste of the reviewer's time:
    They shouldn't even need to make such comments, especially not many
    times for a single patch (see your mention of "overuse"). I realize
    we still have no automated mechanism to check style aspects, but
    anybody can easily look over their patches before submitting them.
    And for an occasional issue I think a terse reply is quite reasonable
    to have.

At the end of the day, none if this is mandatory. The document also
has two audiences
* Authors which get review feedback : for example by just having
this section in there it helps 

I added this section primarily because we do see the occasional
very terse review style and even I think sometimes: wow, that comes
across as harsh. But I also know, that it isn't intentional and that
I have a fairly thick skin. And it is not exclusive to typos and minor issues.

What I was trying to do in this document is to provide
a guide which shows the different patterns from both perspectives.
I hope I succeeded in this, but I believe that you primarily
reviewed the document from the view point of a code reviewer.
    Overall I'm seeing the good intentions of this document, yet I'd still
    vote at least -1 on it if it came to a vote. Following even just a
    fair part of it is a considerable extra amount of time to invest in
    reviews, when we already have a severe reviewing bottleneck. If I have
    to judge between doing a bad (stylistically according to this doc, not
    technically) review or none at all (because of time constraints), I'd
    favor the former. Unless of course I'm asked to stop doing so, in
    which case I'd expect whoever asks to arrange for the reviews to be
    done by someone else in due course.

First of all: this would be our gold standard and as pointed out earlier
So it is intended to provide the tools to do better: for example, from 
my point of view if you followed some of it for example for newcomers
and sparingly when you feel it is right, that would already be a 
win-win. Also, consider that a more positive tone should also have the
effect that there may be less unnecessary discussion. I think this
is particularly true when it comes to the sections on fact-based 
responses vs. some which are unclear. Unfortunately, I don't have data
on this to prove it.
Can I maybe get you to reconsider and re-review the next version from the
view point of an author and maybe make suggestions on how to create more

    I'm sorry for (likely) sounding destructive here.

I don't see this your feedback as destructive and do hope that I
can convince you that documenting some of the patterns which
happen on the list are in fact a net-positive


MirageOS-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.