[FFmpeg-devel] [RFC] replace some static with asm_visibility or so

[Rich Felker]

On Tue, Jan 29, 2008 at 05:46:31AM +0100, Michael Niedermayer wrote:
> On Mon, Jan 28, 2008 at 11:22:04PM -0500, Rich Felker wrote:
> > On Tue, Jan 29, 2008 at 04:46:31AM +0100, Michael Niedermayer wrote:
> > > > > * not following democratic vote on mplayer-dev-eng about uotis account
> > > > 
> > > > Democracy is more than just voting, you need a constitution.  You could
> > > > say that the policy document is such a thing, but it says nothing about
> > > > voting or when an account should or will be revoked.
> > > 
> > > You dont need a constitution, a constitution can help clarify some things
> > > but its not strictly essential. Any group of N people can just vote about
> > > something without writing a constitution first.
> > 
> > I disagree strongly with this claim. Without a constitution specifying
> > how the voting works, voting cannot be meaningful unless there's
> > always a clear dichotomy between choices and everyone agrees on who
> > gets to vote, etc. Whenever there are more than 2 choices, voting is a
> > highly nontrivial matter, and usually when there are only 2 choices
> > it's because the person calling the vote framed the question in a
> > biased way by setting up false dichotomies... (I'm talking about
> > voting in general not ffmpeg/mplayer history...)
> 
> for us:
> who= all developers
> how= well just use debians system if there are more than 2 choices
> 
> also, off topic but IMHO framed voting is better than no voting
> though yes i fully agree there are some cases where it really fails
> like if you can vote between 2 political parties both  having a similarly 
> stupid agenda

keep in mind that even proposing a particular yes/no topic for vote
(like kicking out uau) at an apt moment is a form of framing in
itself. the same sort of thing happens in politics all the time. a
politician will get screwed over one scandal when hundreds of others
walk free after doing much worse things, just because a particular
person in the media or an opposing political faction set them up for
it.

> > These sorts of issues, among other things, are why consensus process
> > is generally more respected than 'plain democratic vote' among folks
> > working in the social justice field.
> 
> the problem with "consensus" is that you do not solve any of the inherit
> problems of a democratic vote. Instead you add more problems.
> 
> consensus amongth whom vs. who votes
> 
> consensus for > 2 options vs. vote for >2
> If you listen carefully, then with consensus stuff everyone just keeps
> trying an option until a large enough majority is reached, and in the
> process people get over time more willing to choose options they consider
> bad
> I would not consider such an ad hoc process optimal, a clear system
> like debians is much more resistant to social manipulation. People
> being very good at arguing, people who happen to be not around while
> the consensus is formed. A vote can easily collect data from people
> over a month not randomizing based on todays available people

consensus is more about recognizing everyone's requirements and
working on a solution that doesn't seriously step on anyone's, as
opposed to just seeking a majority and saying "fuck the rest". success
of course requires everyone to believe in the system to some extent,
i.e. the participants need to want to reach a mutually acceptable
solution (rather than just wanting to "win") and need to be willing to
actively "step up" and "step back" according to the relative degrees
to which each person's needs have been represented so far. it's
obviously not very feasible for a group full of mutual hostilities,
but for one with common goals and vision but different ideas for how
to best reach them, it can be a very valuable tool.

rich