Re: Informal Discussion: Identities of Voters Casting a Particular Ballot are No Longer Public

[Sam Hartman <hartmans@debian.org>]

>>>>> "Don" == Don Armstrong <don@debian.org> writes:

    Don> Without minimizing the totally unacceptable harassment that
    Don> occurred, all three of you seconded the proposals and were
    Don> significantly more visible than a voter listed on the tally
    Don> page.

My suspicion is that if Debian had made a statement in either direction,
this would have expanded well beyond people who seconded the proposals.
I also recall people who became aware of the harassment people who
seconded faced and concluded they were uncomfortable voting.
Which in and of itself is a problem.

But yes, I do understand the point you've been making.


I think there are problematic uses of votes well beyond harassment
though.

* After this, I think the next vote is going to be about firmware.
Do we want companies like Nvidia who may have opinions about how
distributions should think about freedom looking at how people vote when
they consider hiring DDs?

* Do we want ftpmaster members looking back at past votes on firmware
  and DFSG interpretations before deciding someone is an appropriate
  candidate?

* Would it be reasonable for the DPL to look back at votes to decide
  whether to delegate to someone?

* I personally think some of the options on the systemd ballot were
  really bad ideas.  Not just that I disagreed with them, but I think
  that going down that route would have been amazingly bad for teh
  project.
  Is it reasonable for me to  go around holding it against people who
  ranked those options above FD?

I think that if we continue to have public ballots we accept that all
these sorts of things will happen, even if we wish they wouldn't.
The more I think about it, the more I think it would be reasonable to be
concerned about this sort of history building up and to be reluctant to
vote in some cases.
I think for a number of reasons it would be better if we discourage all
the use of voting data I hypothesize above.
I think the best way to do that is not to make the data public and not
to retain it for a long period of time at all.

Personally, I don't think that restricting voting data to DDs solves
many of the issues above.  We could attach a strong policy for how the
data would be used, but by the time we got done figuring out such a
policy, I think we'd just be better off not making the data available in
the first place.

--Sam