[haiku-development] Re: Voting procedure (was: Checking consistency of used strings)

  • From: Niels Reedijk <niels.reedijk@xxxxxxxxx>
  • To: haiku-development@xxxxxxxxxxxxx
  • Date: Wed, 16 Dec 2009 10:43:43 +0100

Hey all,

I hate email for discussions :-) Here a selection of responses to issues raised.

2009/12/15 Ingo Weinhold <ingo_weinhold@xxxxxx>:
> I don't see how you imagine this to work. No consensus was reached in the
> discussion -- there were several versions of title case and sentence case
> favored by different people (I'm actually a bit surprised by the current
> state of the vote) and the discussion eventually died. So what was supposed
> to happen? Someone would just decide that it's option X and define it to be
> the resolution, or what?

Here's the summary: Humdinger decides to make the strings more
consistent. He is willing to invest time in that. He puts some
guidelines up for input, and people have a discussion where two
different viewpoints are there.

At that point I would say that Humdinger can decide from the
discussion which solution has his preference. I imagine (secretly) he
already has a preference. I am worried that he still feels the need to
call for a vote. My viewpoint is that as he does the work, he can
decide what to do.

That's how it has worked, that is how most of the things around here
are working. The one doing the work decides.

2009/12/15 Ingo Weinhold <ingo_weinhold@xxxxxx>:
> Discussion should come first, and if a consensus can be reached: great! But I
> think it's naive to assume that this is always possible.

Maybe I should distinguish between the opinion vote and the 'official'
vote. The opinion vote is when there are two possibilities and the
'worker' cannot decide which one is best. In my view, a developer may
still chose to implement it in way X instead of the more voted for Y,
if he thinks the arguments for X are better.

The official vote is for voting in developers (see other thread
though) and for issues like the alpha 1 requirement vote - choices
that do not necessarily influence only one person's work. Now how
often did we have releases the past few years ;-)

2009/12/15 Ingo Weinhold <ingo_weinhold@xxxxxx>:
> I can't imagine what problem you see. People not interesting in the issue and
> thus not participating in the discussion will likely not vote either.
> Furthermore, if the one doing the work is not happy with the result of the
> vote, than she doesn't have to do the work.

2009/12/16 Axel Dörfler <axeld@xxxxxxxxxxxxxxxx>:
> This is non-sense IMO: personally, I don't vote when I'm either not
> interested in the outcome (why should I vote if I don't care?), or,
> since we hold public votes, if the result is already what I have voted
> for, anyway.
> Why cast 64+ votes, if 5+ votes already do it (and more efficiently
> so)?

Okay, two responses. First of all, if 5 people vote, and 59 find the
issue not important enough to vote on, can you still speak about a
majority decision? Technically yes, but I would say the majority
opinion is 'I don't care, do it your way.' Secondly, the logic that
people that don't vote follow the majority opinion is false. I'm
pretty sure that people who had a stake in the issue are far more
likely to vote than people that didn't. That leads to the fact that a
minority gets to decide.

And with either logic, a 7,8% response rate is pretty bad.

2009/12/15 PulkoMandy <pulkomandy@xxxxxxxxx>:
> For everything else it's quite different : the vote happens when someone is
> not sure what to do. The problem with mail discussion, as we all know, is
> that they tend to get to the bikeshed color and run around for ever or just
> die for lack of answers. It's not like a real-life meeting where you can
> discuss ideas realtime and make a final decision. The vote is then a way to
> end the discussion and is made after some kind of concensus was reached, or
> at least everyone said what he had to. As you can see now in the "sentance
> case" topic, all votes seem to be +1 to sentance case. I think it's quite
> unusual to get a 50/50 result, because of the discussion.
> The problem is that everyone feels the discussion is already ended and there
> is no more need to vote as anyway, everyone agrees.

> The vote still confirms the discussion is ended. It feels better to start
> working on a new feature once you get the majority of +1 instead of working

This is one of the thing I would like to challenge. Unlike Linux who
is one final gatekeeper, Haiku has 64. Every one of those has the
trust to make decisions. I understand that sometimes you would like to
hear some opinions (or in the unlikely event that you are don't know
perhaps even a vote), but every developer should have the freedom to
make his own choices. Setting up an official voting procedure limits
that freedom.

No Haiku developer should feel insecure about patches that he is going
to commit.

If there are obvious reasons why a patch or an approach is bad, it can
always be corrected afterwards on the svn mailing list. That hardly
ever happens...


Other related posts: