Date: Wed, 5 Feb 2003 00:26:15 +0100 (CET) From: =?ISO-8859-1?Q?Magnus_B=E4ckstr=F6m?= <b@etek.chalmers.se> To: freebsd-chat@FreeBSD.ORG Subject: Re: dillon@'s commit bit: I object Message-ID: <Pine.BSF.4.44.0302050025360.82664-100000@scrooge.etek.chalmers.se> In-Reply-To: <200302041820.52497.paul@iconoplex.co.uk>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, 4 Feb 2003, Paul Robinson wrote:
> On Tuesday 04 February 2003 5:56 pm, Magnus B{ckstr|m wrote:
>
> > I would consider a more open governance to be harmful.
>
> Let me guess, you're South American, right? :-)
Nope. Swedish (to the extent absolutely necessary)
> > Things would
> > get bogged down in discussions fed by the opinions of the whole
>...
> OK, sorry for jumping in here (it's been a while since I last posted to -chat,
> but I can't help myself on this one), surely there is a major difference
> between "closed" and "focused". For example, we have many technical mailing
> lists that concern themselves with specific areas of use or development.
> -ports is for the discussion of ports. -small is for the discussion of
> embedded system use. -config, suprisingly enough, is about config and
> installation issues. Anybody in the world who has an interest in those areas
> is free to subscribe to the list and contribute to the project if they so
> wish.
That's focused, on areas of technology. Following you so far.
>
> -core should be no different. If they want to talk in secret, fine they can
> e-mail each other direct, but the only people who are going to be interested
> in reading (never mind writing to) -core are those who have an interest in
> general project management issues surrounding FBSD.
Problem: Focused yes, on areas of technology -- nope.
On a -core list would travel matters of management and policy as well.
Decisions on the latter are ultimately made on principles, on the former
decisions are made on mathematics (in the wide sense.)
The "mathematics" arguments as a rule converge quickly, the "principle"
arguments converge at all only if the argument is confined to a group of
people who are familiar with eachothers' principles.
This is why good management publishes decisions together with a statement
about the principle applied -- it's much, much harder to question the
principle finally applied than to attack a decision in the making by
throwing in foreign principles.
Our acquaintance Brett Glass is skilled at offering arguments that
are plausible along some principle without disclosing which.
The -core list would be a flypaper for Brett Glasses.
>I'd be willing to bet
> that the majority of commiters would not bother subbing to an open -core
> list. Ultimately, the descisions taken by -core not only have an effect on
> commiters, but to people who use FBSD at their place of work and at home. I
> would love certain features to be fast tracked in, but I know that if core
> says "nope, that guy can't be a commiter" they won't be - all I want is a bit
> more of an explanation than "because we say so".
We're entitled to that further explanation.
[But -- aside -- not necessarily in the case of Matt Dillon. The issue
with him is whether or not he has commit privileges, a matter exclusively
for him and his fellow committers. Those who aren't committers ideally
don't have the faintest idea of what problems or blessing are attached
to being a fellow commiter with Matt Dillon, and those who do have a
commit bit are (as I understand) fully informed. End of aside]
> > African tribe
> > management (whole village meets until consensus is reached) may be
> > great as a social event but is disastrous as a direction-finding
> > device.
>
> It seems to have worked quite nicely for African tribes for a few millenia.
> Ahh, but they don't have PDAs and Nike trainers, so perhaps you're right -
> they're obviously not as advanced as us, and always making the wrong
> descisions. Yes, open low-cost democracy where everybody has a voice, quite
> clearly doesn't work. Best off we elect some people who make descision on our
> behalf in secret - a bit like the majority of UK and European parliamentary
> proceedings. They always do a better job, don't they.
I'm not racking down on or belittling African tribes or their culture; I
should have said that the direction-finding device that suits them well
doesn't suit the FreeBSD project very well. For an illustration, imagine
every decision within the project being made by everyone involved --
including you and me -- gathering at a gargantuan BSDCon and not leaving
the place until every decision has been made to everyone's satisfaction.
What you get with more than a minimum of openness in policy discussions
is a tasting of the illustrated chaos and very few benefits. It is
better to appoint a group of people whose principles you generally
agree with, and tell that group "make all the hard decisions, motivate
them, we trust you until our eyebrows rise".
The parliaments you name suffer from obfuscation, which is a real danger
if the group you elect can hide personal agendas in management complexity.
> > The rest of the universe's population gets
> > FreeBSD for free, and can offer clever suggestions but never credibly
> > dictate anything. Not a democracy, sorry -- ergo, not an open government.
>
> So, if you aren't a commiter, you don't add anything to FBSD? Even if I use
> it, write software for it, install it on client sites, train people on it,
> advocate it like god knows what, turn up to confs, encourage people to try it
> until I'm blue in the face, help people (sometimes not very well) in
> arranging user group meetings and confs, I have nothing to contribute to
> FBSD? That, is a ridiculous argument, and one that makes the daemon-bashing
> penguin-lovers of this world seem to have a good point. However, if that is
> genuinely the way -core and many commiters feel about it, I'll be happy to
> unsub from everything at majordomo@freebsd.org and go and find another OS to
> play with.
You have a partly valid argument, but the lists you consider unsubscribing
from is your very defacto instrument for exercising power over the
project. Your clever suggestions indeed get read. Some clever suggestions
take the form of PRs, which -if reasonable- get committed. Some producers
of clever suggestions have such output that the receiving end gets fed
up with being spammed and punishes the source with -- a commit bit, fear.
What advantage, besides a popular predisposition (hype) do the
penguins have? Name an instrument that in essence is different from
anything that FreeBSD has.
> > inside the box. What I see doesn't scare me, and some things are
> > reassuring -- e g the unanimous (trusting greg's statement) decision
> > in -core lately on a managerial matter that is -not- easy.
>
> I've made comments in the past that sometimes when a group decide to keep lots
> of things secret and then just "announce" something, it is often a good idea
> to listen to them. My argument was based on activities of various
> intelligence agencies that don't announce anything without VERY good reason.
> It does not apply to an open source project that is attempting to give the
> impression of openness to all.
The "secret group" is fine as long as their announcement is accompanied
by a principle on which the announced decision was made. If their
decision doesn't follow the principle, they haven't done their job and
should be sent back in. If the decision checks with the principle but
the -principle- meets violent disagreement, the group should be kicked
out and a new one voted in. People who are appointed with the confidence
of their friends (as with -core and the developers) have -very- good
reason to do their job well and to keep their principles clear.
One might argue that -core should be elected democratically by everyone
who has a stake. How then do you measure the stakes, and how do you
-not- alienate those valuable people who "live and breath FreeBSD 24
hours a day" and might be less inclined to put up with the whims of
a group they don't know?
Magnus
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-chat" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.44.0302050025360.82664-100000>