[OpenStack Foundation] Foundation help running a developer survey?
robertc at robertcollins.net
Mon Oct 12 01:18:14 UTC 2015
On 8 October 2015 at 03:48, Barrett, Carol L <carol.l.barrett at intel.com> wrote:
> Robert - Can you say a bit about the information you'd like to gather and the type of actions you're looking to inform from the survey?
was the thread I started in April.
I had some private responses too offering more questions (I've copied
the suggestions, not the authors name as I don't know why they chose
Broadly, I want to be able to get away from squeaky-wheel discussions
when it comes to talking about our contributor base and instead have
some readily available empirical data, so that we're not speculating
about the likely impact of policy changes.
Its likely that with some care the survey could be useful for
academics doing research as well (see below for one set of such
questions), but that philanthropic angle isn't much of a driver for
me. On the other hand, I'd rather add a couple of questions to *one*
yearly developer survey, than have academics from all over doing
de-novo questionnaires to all committers etc and leading to burnout.
For instance, if we find that many of our devs are new to Python, I'd
be advocating the creation of a mentoring program to take
new-Python-developer-mentoring load off of core reviewers. (This is
one of the mild complaints I hear regularly from core devs - that they
see a lot of novice-written-Python and that the education therein is
time consuming - but is it really common, or is it a form of
My draft questions
How many years have you been contributing to OpenStack
[needed to control for the next question]
How many years have you been writing in Python
How fluent do you consider yourself to be in:
How much time do you spend doing:
What operating system do you use to do your OpenStack development
- Are you coding on OpenStack as part of your daytime job
- What percentage of your time can you dedicate to OpenStack
- Is your employer enforcing a policy for upstream contributions
(internal code review, require approval, etc)
I would love to learn a bit about how DevStack is used outside the gate:
* Do you use DevStack as part of your everyday workflow?
* If so, on what distribution(s)?
* Do you run a non-default configuration WRT system services? ie, qpid
or zmq, or psql
* Do you run a gate-like environment using devstack-gate or something like it?
* Do you regularly run a forked/branched DevStack
* Do you run Grenade as part of your local workflow?
I think it would be interesting to also ask about some of the other
tools developed in our community, like gertty. Knowing the usefulness
and adoption of these tools can help justify (or not) ongoing work in
Jesus M. Gonzalez-Barahona asked that the questions they used in their
2014 paper be included to permit updating their effort model. Those
• (1, Selection): On average, how many hours in a week have you spent
in the project in the last six months?
(>40h, 40h, 30h, 20h, 10h, <5h)
• (1, Selection): How much of the time you spent in the project is
devoted to coding? (>95%, approx. 75%,
approx. 50%, approx. 25%, <10%)
• (1, Selection): Do you make at least one commit to the repository
the days you code? (yes, no)
• (2, Selection): What do you consider yourself in the project?
(full-time, part-time, occasional contributor)
• (2, Free-text box): Did you always work on the project the same
amount of hours, or did you have different phases of commitment? If
you had different phases, could you tell us about the various phases?
(the graph below may help you, as it is based in your recorded
activity in the repository)
Can we get some (optional!) diversity questions included? Gender (M/F/
nonbinary), ethnicity, age, # of years experience etc, geography. It's
important we monitor our aggregate metrics here. Maybe also an
agree/disagree on "Diversity is important to the OpenStack community"?
%age of OpenStack time spent for work vs play? Employment status?
Compensation? Might make a persuasive "get into this" argument if we're
compensated relatively highly. Remote working?
"(Launchpad/Gerrit/the specs process/the release process) fulfill their
stated missions" on a scale of strongly disagree to strongly agree.
Robert Collins <rbtcollins at hp.com>
HP Converged Cloud
More information about the Foundation