# Re: GRs, irrelevant amendments, and insincere voting

```On Sun, Nov 02, 2003 at 10:22:50PM -0600, Steve Langasek wrote:
> > Why do you think that voting for "Remove non-free" means that we wouldn't
> > continue to produce a distribution? Why do you think that ballot would be
> > treated differently to:
> > 	[   ] Remove non-free?
> > 	[   ] Don't change
> > 	[   ] Further discussion
> > ?
> Because in your example, the ballot options are non-orthogonal. In my
> example, they are quite deliberately orthogonal,

That may have been your intent, but I don't think you succeeded. Let me go
into a maths analogy to (hopefully) make the point more clear. Consider
a set of axes, we'll call them "x" and "y". Our origin is (0,0), and our
options are to move along the x axis -- ie, some multiple of (1,0) -- or
to move along the y axis -- ie, a multiple of (0,1) -- or to do both. The
vectors (1,0) and (0,1) form an independent basis for this set: which
means any possibility you want can be obtained by multiplying them by
something, then adding them togther. Want (5,6)? Take 5(1,0) and 6(0,1),
and you're done. Other bases are possible: (1,2), (2,1) is another basis,
eg. (0,1) and (0,2) isn't since you can't get (1,0) from it.

The orthogonal axes you're considering are "remove non-free" and "drop
support for various architectures". Possible values of the former are
"don't change", "drop it from the social contract, but keep it in the
archive", "drop it from the social contract and the archive". Possible
values for the latter are "don't change", "drop i386", "drop everything".

But the options above are: (3,0) and (0,0). That doesn't form a basis
for two orthogonal issues: there's no options on the second axis. All
you get is a single dimension -- whether or not to remove non-free;
i386 support will be kept in both cases.

Where's our origin? It's at "keep doing what we're doing now".

> and don't belong on the same ballot together at all

But that's not the case. The option to not change anything at all,
and the option to remove non-free do belong on the same ballot, because
they're on the same axis. Certainly (3,0) and (0,1) mightn't belong on
the same ballot, but that would look like:

[   ] Remove non-free from the social contract
[   ] Drop support for i386
[   ] Further discussion

> -- an attack that the current system has no built-in protections against.

And that's just ignorant FUD. The constitution *specifically* considers
this case, and says that only related options need be considered on a
single ballot.

Cheers,
aj

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

Australian DMCA (the Digital Agenda Amendments) Under Review!
```

Attachment: pgpsg8lO6nCCX.pgp
Description: PGP signature