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

Re: Supermajority requirement off-by-one error, and TC chairmanship

On Thu, Jan 31, 2008 at 07:24:50PM +0000, Ian Jackson wrote:
> The Technical Committee (and those interested in the libc's resolver
> behaviour) are having some trouble because of an off-by-one error in
> the supermajority specification in recent versions of the
> constitution.
> This was discussed in
>   http://lists.debian.org/debian-ctte/2004/05/msg00027.html
> and has recently caused trouble for an actual vote.
> To give a clear and simple hypothetical example: suppose 120
> developers vote on constitutional amendment GR, with a simple Y vs. FD
> ballot (requires 3:1).  Suppose the quorum is met and 90 vote in
> favour and 30 against.
> Then according to the current wording, Y fails to defeat FD by 3:1
> because 90 is exactly 3 times 30, whereas the requirement from A.6 is
> that it should be strictly greater.

I'm not so sure this is an off-by-one *error*; for example, when simple
majority is required, then a strict 50% against vs 50% in favour result
should result in the status quo being kept. A simple majority thus needs
to say "*more* than X percent of people need to agree for the vote to
pass", and it seems reasonable to give a supermajority system the same
set of rules.

Having said that, I agree with you that it makes sense for the TC to not
require 'X + 1', since the electorate is so small anyway; and on the
whole of Debian, the numbers don't really matter that much, anyway; in
principle, I don't object to this change for supermajority votes. In
effect, it's a change of what "supermajority" means, but with not much
problems in practice (and serious practical benefits, as in the TC).

> I suggest the following wording:
>  * Replace `strictly greater than' with `at least' in A.6(3)(2).
>    The result reads:
>        2. An option A defeats the default option D by a
>           majority ratio N, if V(A,D) is at least N * V(D,A).
>    The immediate effect of this is that options which are tied with
>    the default option are not dropped;

I don't think that's a good idea. If a vote is tied with the default
option, it failed to convince a *majority* of voters. I don't think the
principle of majority is one that should be touched if we want to hold
the semblance of democracy in our votes; for that reason, I can't agree
with your proposal as is.

If you were to change it so that your proposed change would only apply
to supermajorities, I'll probably second it.

>    options with a supermajority requirement are likewise not dropped
>    if they exactly meet the supermajority requirement.
>    This has two practical consequences:
>    Firstly, the off-by-one error in supermajorities is fixed: if the
>    number of Yes voters is exactly 3x or 2x (as the case may be)
>    the number of FD voters, the resolution is considered successful.
>    (In a vote with no cycles and where a position option meets a
>    supermajority, FD will be eliminated from the Schwartz set by
>    A.6(6).)
>    Secondly, where there is a casting vote, it allows the elector with
>    the casting vote to choose between FD and Yes if they are tied.

The rest, I think, you should put in different proposals; I have no
problem with them, though, and could probably second them.

> While we're looking at this text, the use of `defeat' is slightly
> confusing, and made more so by this change, because the criterion for
> defeating the default option in the new A.6(3) is different from the
> definition of defeat provided in A.6(4).  So I propose one other
> change to go along with this:
>  * Change to `defeat [the default option]' throughout A.6(3) to
>    `overcome'.  That gives:
>       Any (non-default) option which does not overcome the default
>       option by its required majority ratio is dropped from
>       consideration.
>        1. Given two options A and B, V(A,B) is the number of voters
>           who prefer option A over option B.
>        2. An option A overcomes the default option D by a majority
>           ratio N, if V(A,D) is at least N * V(D,A).
>        3. If a supermajority of S:1 is required for A, its majority
>           ratio is S; otherwise, its majority ratio is 1.
> Additionally, the TC had considered experimenting with the idea of
> rotating the post of TC Chairman, to try to make things a bit more
> dynamic.  However it's not clear with the current constititution that
> the TC has the power to (for example) set out a timetable specifying
> the chairman at various times.
> So I would like to suggest replacing 6.1(7) with something like:
>   7. Appoint the Chairman of the Technical Committee
>      The Committee may by resolution specify which of its members is
>      its Chairman.  If it does not do so, or circumstances arise which
>      are incompatible with the resolution, the Chairmanship is vacant
>      until this situation is remedied.
>      /Rationale: For example, the Committee might establish a schedule
>      for a rotating Chairman, or appoint a Deputy to take over under
>      specified conditions./

<Lo-lan-do> Home is where you have to wash the dishes.
  -- #debian-devel, Freenode, 2004-09-22

Attachment: signature.asc
Description: Digital signature

Reply to: