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

Re: Debian Etch Stable.

On Thu, Dec 14, 2006 at 03:32:16PM +1000, Anthony Towns wrote:
> On Wed, Dec 13, 2006 at 01:24:32PM +0000, MJ Ray wrote:
> > > Actually, I believe you'll find that that wasn't even put forward as a
> > > metric for the experiment.
> > In your own words, the experiment was to allocate sufficient funds so
> > that Steve Langasek and Andreas Barth can dedicate a month each to
> > getting etch out on time (and Mon 4 Dec 06 was already given as the
> > release date).
> If you consider that to be the "success condition", it seems it was
> already a "success" -- that amount of funds was allocated, for exactly
> that purpose.

I've said a few times in this discussion things like "as long as you
learn something, an experiment isn't a failure". If you combine that
with the fact that when you learn something you do things differently,
that's equivalent to saying that if the experiment needs to be repeated,
it's a failure, and if it's a success, it doesn't need to be repeated.

I'm going to continue avoiding trying to draw any conclusions about the
effect on Debian while the release is still pending, but at the very least
there's been a few things learnt:

	- Debian shouldn't do funding of Debian work for various reasons,
	  that add up to it not having the support of the developer body

	- there's no need to do funding through Debian/SPI -- it's possible
	  to raise decent amounts of money for Debian work in short amounts of

	- lots of people don't like key Debian people to be involved
	  in funding Debian things, due to potential or actual confusion
	  and other potential or actual conflicts of interest

	- some people don't like paying people at all and would rather
  	  that not affect their hobby

	- some people actively disendorse/disclaim the goal of releasing
	  on a set date, even as subordinated to "when it's ready"

A number of those were found out during discussion on -private, and
Dunc-Tank aimed to address them as best it could; posts to this list
and elsewhere demonstrate that any future projects should be able to do
a better job of that IMO. I'm pretty sure there're other lessons that
could be drawn out of a better understanding of the objections related
to how the task was selected and whatever too. No doubt some of these
conclusions were obvious to some people before this was ever raised;
unfortunately not everyone's that well informed.

So as far as I can see, questions like "Has the "experiment" failed
now? Will it be repeated?" [0] are related in the opposite way to what
people are assuming -- yes, things were learnt, therefore no it hasn't
failed, and no, the experiment doesn't need to be repeated [1].

I presume the real question is whether anyone gets funded to do Debian
stuff in the future. But ultimately I don't get /any/ say in that --
it's totally up to the people who are actually going to do the funding.
If you're against this sort of thing, it's not me you've got to convince
of the evils of funding or the superiority of other ways of doing this;
it's all the people who can afford to fund Debian-related activities,
and have the desire to do so. Because in the end, they'll do whatever
they want with their money, no matter whether you, or I, or Debian thinks
it's good or bad, or whether we dance for joy, or sulk and swear.

(As an aside, one question that has been at the back of my mind following
some of the more intense opposition is just how easy it might be to
disrupt Debian by applying small amounts of money -- either to recruit
developers to work on other projects, or just to create arguments that
distract people from doing work (ie, "trolling"). If it's really possible
to repeatedly disrupt Debian with just $12,000, eg, I'd imagine Microsoft
would be happy to do that in a heartbeat if they ever started feeling
threatened by us. Again, I don't want to judge things 'til etch is out,
but my initial impression is that in spite of the flamewars and claims of
demotivation, the primary "negative" effect has been an increased focus
on quality assurance (in support of the widely recognised goal of making
the next stable release as reliable and well-tested as possible), at
the expense of releasing on the 4th (which was a less expected goal, and
less widely recognised amongst developers). If that's actually the case,
that's a fine kick in the pants to me or Dunc-Tank supporters or whatever,
but at the same time it's a good demonstration that Debian's willing
and able to keep its own priorities in the event of monetary interference)


[0] http://www.infodrom.org/~joey/log/?200612190920

[1] ...at least if you're not trying to get published in an academic
    review and have to demonstrate repeatability under laboratory
    conditions or similar.

Attachment: signature.asc
Description: Digital signature

Reply to: