[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: GRs, irrelevant amendments, and insincere voting

On Sun, Nov 02, 2003 at 10:28:33AM -0500, Raul Miller wrote:
> > On Nov 2, 2003, at 02:45, Raul Miller wrote:
> > > How is an amendment appearing on the ballot equivalent to a veto?

> On Sun, Nov 02, 2003 at 09:31:20AM -0500, Anthony DeRobertis wrote:
> > Because our voting system can only provide one winner, even when the 
> > options are orthogonal. So, a very popular option line "keep x86" 
> > effectively veto's a less popular (but still 3:1 supermajority) option.

> > Since our system can only provide 1 winner, I think that orthogonal 
> > options should not be allowed on the same ballot.

> In other words:

> [1] if the proposer of some ballot option chooses to ignore some popular
> amendment

"Popular" only in the sense that it expresses a view that is popular --
not that the idea of replacing the ballot option with the amendment
receives popular support.

> [2] (and chooses not to provide an option which includes the most salient
> points of both),

Consider the "amendment" (in name only),

   Replace lines ^ through $ with the words, "Debian should continue to
   produce a distribution."

Such an amendment would only exist to subvert the original purpose of
the ballot, so there is no way to incorporate elements of it into the
original proposer's ballot option.  In many procedural systems, this
would nevertheless be considered an amendment, and our SRP does not
*systemically* prevent such an interpretation.

If people vote sincerely (i.e., they vote in accordance with how much
they truly agree with each ballot option), this particular option would
stand a good chance of beating out *any* other option it was put up
against before the developership.  Or, developers may see through this
ruse, and vote *against* the amendment; in which case there is an
element of insincerity to the votes they've cast -- they are voting
strategically, rather than according to their true preferences.  I
believe this is undesirable.

> it's possible that that amendment will be more popular than the original
> proposal, even though

> [3] that popular amendment doesn't include something the proposer of
> the original option considers salient (and might, in a separate ballot,
> get approval for).

> I agree that this is the case.

> I don't agree that this is a flaw in the voting system.

> If people want to play games, rather than work directly towards the
> best outcome, then the result will be an indirect approach towards the
> best outcome.

And the Project Secretary has indicated he would use his Constitutional
power to prevent orthogonal ballot options from being voted on together,
to thwart any such attempts to subvert the system.  Which reduces the
scope of this vulnerability to "How you, the Project Secretary, and four
of your friends can kill any GR." :)

Consider also that Debian is now a rather high-profile user of a
progressive voting system and may well find itself used as a model by
others.  Even if you believe Debian will never have an unreasonable
Project Secretary, there may be some larger benefits to identifying and
eliminating any formal procedural flaws in our voting system.

Steve Langasek
postmodern programmer

Attachment: signature.asc
Description: Digital signature

Reply to: