[openstack-dev] [all] A proposal to separate the design summit

Jeremy Stanley fungi at yuggoth.org
Wed Mar 2 18:05:21 UTC 2016


On 2016-02-29 15:03:19 -0800 (-0800), James Bottomley wrote:
[...]
> it sounds like an an expectation that people who aren't gamers
> would submit more than one patch and, indeed, become part of the
> developer base. I wanted to explain why there's a significant set
> of people who legitimately only submit a single patch and who
> won't really ever become developers.

Some rough curve-fitting was performed based off per-contributor
patch counts over a cycle and the observation was that single-patch
contributors were a significant positive deviation. The model
suggested that we'd have somewhere around 10% fewer active
contributors in a cycle if we calculated the single-patch
contributor projection based on the counts of contributors with an
increasing number of merged patches that cycle. This was used as
justification not to increase the minimum patch requirement for a
free summit pass, since the prediction was that would simply move
the 10% bump to whatever the minimum number of patches was to
qualify.

What I'd love to do, but have yet to find the time, is perform a
more detailed analysis and model incorporating a mapping of which
contributors exercised their free admission. This would provide a
far more accurate picture of whether people are really contributing
just one patch so they can save US$600 on conference admission, or
whether we have a disproportionate number of single-patch
contributors for some other more serious reason. Honestly, if the
rough model turns out to be true, then serving 90% in the way we
intended with only 10% freeloading seems fine to me (as long as
people who have no real intention of contributing stay out of design
sessions and let the rest of us get work done).
-- 
Jeremy Stanley



More information about the OpenStack-dev mailing list