Re: Commitfest problems

* Jaime Casanova (jaime(at)2ndquadrant(dot)com) wrote:
> On Tue, Dec 16, 2014 at 11:32 AM, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
> >
> > It has been proposed that we do a general list of people at the bottom
> > of the release notes who helped review during that cycle. That would
> > be less intrusive and possibly a good idea, but would we credit the
> > people who did a TON of reviewing? Everyone who reviewed even one
> > patch? Somewhere in between? Would committers be excluded because "we
> > just expect them to help" or included because credit is important to
> > established community members too? To what extent would this be
> > duplicative of http://www.postgresql.org/community/contributors/ ?
>
> Not much really, I tried to get my name on that list a couple of years
> ago, when i reviewed more than i do now, and never got it.
> And while my name is in a couple commit messages, that is a lot harder
> to show to people...

Having your name in a list of other names at the bottom of the release
notes page, without any indication of what you helped with, would work
better? Perhaps it would but I tend to doubt it.

If the release notes were expanded to include more than just a list of
reviewer names, and more along the lines of the way patch authors are
credited, that would be more valuable for the reviewers but I'm not
excited about such an increase in the size of the release notes. To
play it out a bit though, we could do something like:

* Add Magic PostgreSQL feature (Author: John Doe, Reviewed By: Jane Doe)

With the commit history now including that information, generally, this
would hopefully not require too much additional work for the release
notes authors. As for the 'rules' about who is considered a reviewer
for these purposes, I'd suggest it just go off of whatever is mentioned
in the commit history but exclude committers (and perhaps major
contributors..? they are already acknowledged on the website, after
all..). That would help keep the lists of reviewers from getting too
large, hopefully.

I'm pretty sure this was brought up previously and shot down but perhaps
feelings have changed about it now, especially as the information is
generally in the commit history and there is already some amount of
"policy" regarding who gets what credit as it's more-or-less up to the
committers. I definitely think that, in general, reviews need to be
more than "it compiles" to get that credit.

> you know, it's kind of frustrating when some not-yet customers ask for
> certificated engineers, and there isn't any official (as in "from
> community") certificate so you need to prove you're a contributor so
> let's see this random commit messages...

Another thought I had was to suggest we consider *everyone* to be a
contributor and implement a way to tie together the mailing list
archives with the commit history and perhaps the commitfest app and make
it searchable and indexed on some website. eg:

contributors.postgresql.org/sfrost
- Recent commits
- Recent commit mentions
- Recent emails to any list
- Recent commitfest app activity
- Recent wiki page updates
...

Ideally with a way for individuals to upload a photo, provide a company
link, etc, similar to what the existing Major Contributors have today.
Obviously, this is not a small task to develop and there is some risk of
abuse (which I expect the other folks on the infra team will point out
and likely tar and feather me for suggesting this at all..) but it might
be along the same lines as Bruce's PgLife..

Thanks,

Stephen