Open Stack

On Wed, Oct 30, 2013 at 10:08:03AM +1100, Tom Fifield wrote:
> Hi all,
> 
> Recently, I did something crazy and got into the "top 10" reviewers
> for OpenStack in a 30/60 day window. Admittedly, this was for
> documentation  - which is quite a bit different than code - but the
> experience did give me a small window of insight into the challenge
> faced by our venerable core reviewers. It's a really tough job!
> 
> One of the aspects that I noticed in doing so many reviews is that a
> review was much easier to perform if another reviewer had been
> through it beforehand. That is, a patch had gone through a couple of
> -1 iterations to finally get a +1 before I saw it.
> 
> This made me think a little about how much emphasis we place as a
> community on +2 reviews. It can seem at times like they're the only
> reviews we care about. Hell, I've even heard song lyrics from a
> community member that imply this :D
> 
> I think it's time to bend that focus slightly, and promote the role
> of the +1 reviewers. Every review that a non-core reviewer does
> helps reduce the burden of core reviewers just that little bit.

It absolutely does, and is much appreciated by us core team
members.
 
> Do you see this too? How can we help encourage more +1 reviews?

It is a tough question. You don't want to put up strict rules since that
is typically counterproductive. Perhaps the biggest carrot to encourage
more +1 reviews, is that it is a stepping stone to becoming a core team
member. eg if you find yourself in the top-10 reviewers on nova for an
extended period of time you'll likely get an invitation to become a
core team member from Russell.

Looking at our wiki page 

  https://wiki.openstack.org/wiki/How_To_Contribute#If_you.27re_a_developer

it is very much focused around that idea that you have to write code or
do code fixes to become involved. It isn't really mentioning contribution
via reviews at all. It merely mentions "learn gerrit" and use it to sign
the CLA.

Similarly this page

  https://wiki.openstack.org/wiki/Gerrit_Workflow

only mentions review in the context of what happens to *your* patch.

I think that both those pages could be extended/re-written to strongly
encourage contributors to participate in reviewing other people's patches.

The "How to Contribute" page should explicitly list "reviewing code
changes" as a primary way to contribute, and mention that it is one
way to get on a path towards gaining the reputation needed to become
a core team reviewer/member.

The "Gerrit Workflow" page should say something like

  "While you are waiting for review feedback on changes you have
   submitted, you should browser other pending changes and provide
   review on them. By helping out with other reviews you will be
   decreasing the time delay for your own change to receive feedback"

Daniel
-- 
|: http://berrange.com      -o-    http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org              -o-             http://virt-manager.org :|
|: http://autobuild.org       -o-         http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org       -o-       http://live.gnome.org/gtk-vnc :|