[OpenStack Foundation] Individual Member Election Statistics

Matt Joyce matt at nycresistor.com
Tue Oct 2 23:09:13 UTC 2012


I'll respond in line.  Since there is a bit more here than a single
simple response.

On Tue, Oct 2, 2012 at 3:06 PM, Stefano Maffulli <stefano at openstack.org> wrote:
> On 10/02/2012 01:29 PM, Matt Joyce wrote:
>> The question that remains however, is...
>>
>>  Are the foundation elections being manipulated?
>
> I would like to understand why you're asking such unfair question.

I don't think it is an unfair question.  The data has some pretty
glaring anomalies which I've taken the time to point out.  I believe
that's acceptable open debate under the community code of conduct, and
as I said I think it's important.  Quantifying our metrics and
measuring what we can is the first step.  Correlating is the second.
If we wish to be able to provide useful recommendations based on the
data released earlier in this thread we'll need a healthy
participatory discussion on what this data can be tied to and whether
or not it is relevant.  This will provide us with a path to provide
formal peer reviewed analysis and recommendations built not on the
data we have and not the data we don't.

> Do you think that people got elected without deserving their seat? If you
> think this is the case, please speak up: you can talk to Jonathan or to
> the Foundation's lawyer.

I assess "deserving their seat" as being a function solely of the
election.  As it should be and is claimed to be.

As stated, the foundation membership requires ALL members to avoid
manipulating the election results.  However the data seems to indicate
that manipulation may have occurred.  If I had specific and DEFINITE
evidence of manipulation I would of course forward that to the
foundation lawyer.  I have been clear in saying I have no such
evidence.  However there is clearly a large number of anomalies in the
election results that would be best explained as being the result of
manipulation ( as far as I can tell ).  The real cause for concern
here is that there is no way to confirm that manipulation of the
election did not occur, and there are unexplained anomalies in the
results that very clearly affected the elections in a serious way.  4
out of 11 of the top results in the election were elected by
co-workers ( greater than 75% of their total votes came from their own
employees), and not by a matched percentage of outside users.  With
three companies having members totally nearly the entirety of the rest
of the membership each, I cannot discount the possibility that
intentional manipulation has occurred.  I just can't.  The numbers are
there.  It's just plain overwhelmingly suspicious.

So I have to ask if the election procedures as they exist today
reflect the foundation users real desires.  Based on the numbers I
have, I suspect that is not the case.  So, do I believe that people
got elected without "deserving" their seats?

Let me respond by saying.

Logically I cannot say definitely that the people elected in the
foundation elections were done so without outside manipulation.  There
are anomalies in the data we have that cause me to believe there is a
very real possibility that the election was manipulated.

However, without proof that the election was manipulated, it would
appear we must accept the results of the election as being valid.
Thus the members elected deserve their seats as far as procedure
dictates ( or my understanding of procedure anyways ).

The fact remains however, that there is no means to verify the
election was not manipulated, and there is evidence that suggests it
may have been.  I am concerned by that and I think open discussion and
debate concerning that issue is valuable to the foundation and the
openstack community.  More to the point it's necessary.  People will
look at these numbers and draw conclusions from them on their own.
Fact is I can't explain the numbers as they exist today without
assuming direct manipulation of the election.

> The thing is that the Foundation was designed to have very little
> barrier to entry for Individual members. With such design, of course
> large corporations will have more members and of course they will vote
> for their colleagues, if nothing else because those are names they
> recognise.

Yes, this has been said before.  However, without a means to verify
that manipulation of the election has not occurred, and with clear
anomalies in the election results that remain unexplained I have to
ask if this was the right call.  As far as I can tell we've provided a
means to game the elections and eliminated any means of verifying the
authenticity of the results.  We don't have a specified threshold of
percentage error we're willing to tolerate here.  This is entirely too
ambiguous for me, and with three companies that do NOT represent the
3/4ths of the global openstack community driving the foundation
membership ( as far as any numbers I've seen can indicate ), I have to
ask if this works at all.  This is pretty damned important and the
numbers don't assuage my concerns, they exacerbate them.

> With such design, the Foundation has put in place checks and balances
> *after* the joining process to avoid giving too much power to Platinum
> sponsors and corporations in general. Like all systems, it can be
> improved but we need to at least make it run a few times before saying
> it's broken.

There's a heck of a lot more that we can do to track this stuff.  I am
pleased as punch to see the data sets and collection utilities we have
seen brought forth by the community.  I think a healthy analysis of
that data, and the collection of more data can be used to help
identify problems in the electoral process.  I think discussion and
debate is a good thing even at this early juncture.  However, at the
same time the severity of the anomalies in out first ever elections
are severe and that DOES taint the election results.

The checks and balances in place today do nothing to limit or even
identify external manipulation of the foundation membership elections.
 I'm all for not solving problems we don't have, but the numbers as
they exist today tell us we have a problem we need to solve.  Direct
manipulation of the foundation membership election is not only
possible, the numbers as they exist today match what a manipulation of
the foundation membership elections would look like had it definitely
occurred.

> I'd like to bring the focus on the people that have been elected, help
> them do the interests of the category they've been elected to represent.
> If they do then we can stop the speculations around OpenStack and the
> power of its corporate members (there are quite a few haters out there
> that do that job nicely already). If they don't, then we have tools in
> the bylaws to remove them.

Personally I think everyone who ran for the director positions was
well qualified and even among the folks I've identified as having
anomalous results I think the folks are exceptionally capable and well
qualified for the director seats.  Heck I think I even voted for rob
hirschfeld ( whom was identified as having anomalous results ).  I
certainly think Monty Taylor ( also anomalous results ) is an amazing
contributor and will do phenomenal work on the board.  I don't think
either of these guys from my past experience would ever seek to
manipulate the elections themselves.  But I have no doubt people in
our industry would take actions that could be construed as direct
manipulation.  I'm not about to call anyone out as being unqualified.
That's not the point of this thread or my post.  Nor is it the goal of
this discussion to assist new directors.  This thread is regarding the
analysis of data returned from foundation election audits.  I want to
make that audit more complete.  I want to be able to identify
suspicious anomalies in our electoral data results.  I want a baseline
that I can correlate to our existing community participation metrics
that can explain the way the voting is occurring. I want to formalize
our approach to analyzing results and then be able to provide
quantifiable metrics to support recommendations for how elections can
happen better.  In that I think you may have missed my goal entirely
as well as the point of releasing this data.

Stefano.  I love OpenStack as much as you do.  And I respect everyone
in this community.  This is not an attack on the community, or the
directors.  This is an honest nothing held back discussion of hard
numbers.  That is all.  I do have opinions, and they color the tenor
of my text.  I do my best to remain objective.  I fail occasionally
and I apologize for that in advance.  But this isn't a bogus issue.

TL;DR summary.

If a company had intended to directly manipulate the foundation
elections and proceeded in the most obvious manner, the election
results demonstrated by Dell in this data set would very nearly match
what you would expect.  You can call that coincidence if you want.
But it's not just Dell.  HP matches it, and Rackspace is right there
next to them sitting in the exact same predictable distribution range.
 I am concerned because every ounce of data we have screams that there
is a clear correlation to direct manipulation of the election and
nothing to explain that correlation is anything but that.

-Matt



More information about the Foundation mailing list