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

Re: Default options (Technical Committee supermajority)

On Thu, Jun 10, 2004 at 01:27:17AM +0100, Ian Jackson wrote:
> 1. Supermajority:
>   TC is never more than 8 people, this effectively means that the TC's
>   supermajority for overruling a developer has been changed from 3:1
>   to 4:1.

So, changing that to a X >= m*FD, means in the m=1 case that you don't
drop an option which is ranked equally to FD -- the other cases have large
majorities anyway, so they're not interesting. In the case of:

	100 XD
	100 DX

the current system says "further discussion wins", adding an "=" would say
"the elector with a casting vote chooses between further discussion and X".
That's probably okay, maybe even desirable.

The general case is when D is in the Scwartz set of undropped options. D
is always undropped -- but in the current system it's only in the Schwartz
set if all other options are dropped. In the new system, every undropped
X either defeats D, or is in a pairwise tie with D (otherwise it would've
been dropped). The interesting case is when D's in the Schwartz set. Hrm.

D can't transitively defeat anything -- it'd have to pairwise defeat
the thing directly (in which case it'd already have been dropped), or
pairwise defeat something else, which transitively defeats the thing we're
looking for (but then that something else would've been dropped). So the
only way for D to be in the Schwartz set is if all the other undropped
options fail to transitively defeat it, too. Which means it has to be
in a pairwise tie with everything.

Some examples:

	100 ABD
	100 DAB

gives you A=D, B=D, A>B. A transitively defeats B, and isn't transitively
defeated by D, so A is in the Schwartz set. B isn't. A doesn't t. defeat
D, nor does B; so D is in too, and the casting voter gets to choose
between A and D (but not B).

	100 ABD
	100 DBA

gives you A=D, B=D, A=B; so the elector with a casting vote decides
the winner amongst all the options.

	50 ABCD
	10 BCAD
	30 DBCA
	30 DCAB

gives you 60:60 A,B,C=D; and 90:30 B>C, 80:40 A>B, 70:50 C>A. Your
Schwartz set starts at {A,B,C,D}, you drop C>A, which reduces your
Schwartz set to {A,D}, and your elector with a casting vote decides
whether A or D should win.

That seems reasonable -- I think changing "strictly greater" to "greater than
or equal" A.6.3.2 should work.

>   This weakens the ability of the TC to overrule a developer when
>   exactly one TC member disagrees.

In the case of supermajorities, the default option wouldn't be in
a pairwise tie with all undropped options, so wouldn't get into the
Schwartz set, so there'd be no other changes there.

> 2. Casting votes:
>   6.3(2) says that the TC chairman (that's me ATM) has a casting vote.
>   However, A.6(3) and A.3(3) mean that the casting vote cannot be used
>   to choose between Further Discussion and another option - FD always
>   wins.
>   This takes away much of the point of the casting vote, which is to
>   allow an issue to be resolved even when the committee is `hung'.

The "casting vote" is there for resolving ties; "further discussion"
is there for ensuring all applicable options have been put on the table
and thought through. If you're going to get a hung result, it should
be between two non-default options. I can't say I'm too bothered by
that result; but I'm not against changing it either.

> The library dependencies vote would have
> resulted in Further Discussion winning.

(at http://lists.debian.org/debian-ctte/2002/07/msg00018.html)

>  Any (non-default) option which does not suffice compared to the
>  default option 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. M:N is the supermajority ratio required for A; M:N is 1:1 if no
>       supermajority is required.
                         ^^^^^^^^ specified 

>    3. An option A suffices compared to
>       the default option D, if M * V(A,D) is equal to or greater than
>       N * V(D,A).

If Debian only has S:1 supermajorities, I don't see the point complicating
things by talking about M:N supermajorities. *shrug*


Anthony Towns <aj@humbug.org.au> <http://azure.humbug.org.au/~aj/>
Don't assume I speak for anyone but myself. GPG signed mail preferred.

``Like the ski resort of girls looking for husbands and husbands looking
  for girls, the situation is not as symmetrical as it might seem.''

Attachment: signature.asc
Description: Digital signature

Reply to: