Re: testing and no release schedule
On 28-Mar-04, 22:24 (CST), Adam McKenna <adam@flounder.net> wrote:
> On Sun, Mar 28, 2004 at 06:40:02PM -0600, Steve Greenland wrote:
> > Sure. But the thing is, I only signed up to maintain a few packages.
>
> Then do that. But do it with the right mindset. Debian is about releasing a
> high quality operating system, not just keeping unstable up to date.
You're missing my point, or I'm not making it very well. My point is
this:
Certain things have to happen for a release. One of them is a getting
rid of RC bugs, and indeed, a maintainer you can't/won't do that for
their own packages is failing their basic commitment. But even then, you
have to choose between dropping those packages or hoping someone else
will come along with an NMU.
Another is getting the installer working. But if no one wants to work on
it, or only a few, then there's nothing "management" can do to hurry it
along. It takes as long as it takes, and the effort put into it is based
on how many capable people volunteer to work on it. There's simply no
structure in Debian that allows the DPL, or anyone, to say "Hey, Fred,
go work on the installer, and don't do anything else until it's done."
That's what people mean when the say "Debian's a volunteer org, and
thus things take as long as they take, and you can't do anything about
it." To simply reply "That's BS", as you did, implies that you don't
understand that, or that there is some conotation that you're replying
to that *I* don't understand.
> >
> > Depends on the user. People with a room full of servers don't, in fact,
> > want frequent stable releases.
>
> I have a question.. Do you actually have a room full of servers? And I'm
> not talking about your bedroom or the Computer Science lab at school. I mean
> a room full of servers that an actual business depends on. I'm not asking
> this question to be derisive, I just want to know where you're coming from.
I, personally (or rather, the company I work for, and whose server
maintenance is one of the things take care of), have several servers
at work that perform the usual business functions (mail, customer DB,
etc.), that only get upgraded if there's a specific reason to do so. If
I had to upgrade them every six months, it would annoy me, and it would
suck time from my main job (developement), but wouldn't be the end of the
world.
My coompany provides software and support services to several fairly
large organizations (banks, some govt orgs, large school districts,
etc.). Some of them feel as you do, and upgrading frequently would not
be a big deal. Others, however, have rigourous policies and requirements
for upgrading such services, and large testbeds in which to validate
*any* change. Upgrading an OS is likely to be a several week effort.
There are fasttracks for security fixes, but only if they are indeed,
security fix only changes, not feature upgrades. These organizations
don't do this just because they are paper-work happy, but because
downtime costs money, and they've decided that spending money validating
changes is cheaper than being down. For them, the slow Debian releases
are not a disadvantage.
> I have a room full of mission critical servers. Several rooms, in fact, and
> I don't feel how you're claiming I feel.
I should have said "some people with a room full of servers...", yes.
>
> > > Developers who state openly that they do not care about releases are
> > > in violation of the Social Contract.
> >
> > Oh, crap. There are *lots* of different kinds of users. If nothing else,
> > every Debian Developer is a user. So if a certain subset of them are
> > happy with unstable, what's the big deal? If all I had to deal with was
> > my home desktop, I'd be happy with unstable, too.
>
> Nothing's wrong with them being happy with unstable. I'm happy with
> unstable. But there is a difference between being happy with unstable and
> publically stating that you do not care about releases.
Why? Assume that Debian developer H is happy with unstable, and keeps
her packages bug free and current. Why should she have to care about
some other group of Debian developers who want to make releases?
I never said (or meant to imply) that anyone was working to sabotage
releases. But there are people for whom they are irrelevant, I don't
think there is necessarily anything wrong with that.
> There's actually a word for a software product that never releases.. It's
> called 'vaporware'. I trust you've heard this term before?
I'm running Debian on the machine I'm writing this on. It's hardly
"vaporware".
> If debian turns into a product that only serves the needs of its
> Developers, and not its users,
The Developers *are* Users. Even if there are only 900 Debian users, if
it's the right 900, then Debian will keep being useful.
Of course, there are *lots* of developers who do care about releases,
and doing things that make it easier/better for non-developers, and so
forth. But there are some who don't, and as long as they don't interfere
with those who do, I don't see the problem.
And that's about all I have to say on this topic (the crowd goes wild!).
Steve
--
Steve Greenland
The irony is that Bill Gates claims to be making a stable operating
system and Linus Torvalds claims to be trying to take over the
world. -- seen on the net
Reply to: