Re: GR Ballot Option: Allow, but do not require, secret voting
On Thu, Mar 3, 2022 at 3:12 AM Judit Foglszinger <urbec@riseup.net> wrote:
> I think, 4K puts the bar very high (that would require 20 people).
On Fri, Mar 4, 2022 at 12:39 PM Bill Blough <bblough@debian.org> wrote:
> However, to generate further discussion, I do agree with Judit [1] that
> 4K seems like a high bar.
Hi Judit, Bill,
I agree, 4K is high. I struggled with this a bit before I proposed
it, and I'm still not 100% sure it's the right balance, but this is
some of my thought process:
Unlike "regular" ballot options, a ballot option containing a vote for
secrecy is a final decision once it reaches the threshold. It does
/not/ require voting to have an effect; rather, the mere existence of
a ballot option requesting secret ballots with 4K seconds will fix the
election to be a secret ballot, regardless of the outcome of the final
vote. Even if zero people voted it above NOTA, it still would turn
the election into a secret ballot.
In this way, it feels to me like it should have a higher bar than what
is regularly required to propose a ballot option (K). Looking at the
constitution, there are two other thresholds dictated about voting:
enough to put a decision on hold (2K = 10), or the quorum required for
a vote to be valid (3Q, approximately 48 people at the time of this
writing).
Because the constitution clearly wants to maintain a minimum threshold
for discussion of ballot options (i.e., setting K as the floor of Q or
5), I didn't want to use Q as a basis for this threshold. It seemed
wrong to set it to 2K, though, considering that 2K developers merely
puts a decision on hold -- the final vote is what determines the
actual outcome. Because this ballot option is itself a mini-vote
inside a vote (as it has an independent effect regardless of the vote
outcome), I decided it should be higher than 2K.
Why 4 and not 3? I don't have a great logic for it. I looked at
recent votes and counted the number of overall seconds between all
proposals, and set it based on that number to be high enough so I
thought it was realistically achievable, though a high bar.
In practice, the way that I would like to see this work is that a
ballot option is proposed with no content other than turning the
ballot to a secret option. Then people can, regardless of their
position on the issue, second that ballot option to avoid splitting
the vote. This would have been easily achievable on the RMS GR (n=47
unique seconds + proposers) and on the systemd GR (n=42 unique seconds
+ proposers), but not on the less controversial declassify debian
private vote (n=18 unique seconds + proposers).
I still hope that this option receives enough seconds to go on the
ballot as an intermediary position between the current options of
"always secret" and "never secret except for DPL elections".
Sincerely,
--
Harlan Lieberman-Berg
~hlieberman
Reply to: