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

Re: next debian stable ?



On Tue, Apr 29, 2003 at 12:22:07PM -0700, nate wrote:
> Noah Meyerhans said:
> 
> > You don't need to be commercial to have goals or a timeline.
> 
> true but debian is run by volunteers. what do you do when one of
> these volunteers(which probably doesn't get paid at all) doesn't
> meet the deadline? and nobody else steps up to the plate?  "Firing"
> the "volunteer" usually doesn't do much good, especially if that
> person is contributing elsewhere. Sometimes the packages just end
> up shipping with bugs(e.g. sessions were broken on many architectures
> in php4 on woody), other bugs are more serious though.

Since I'm usually on this side of the debate, I thought it might be
fun to play both sides for a change.

I used to work in a "professional" software development environment.
I learned a lot about software development :-)  In some ways I was
shocked and disappointed; deadlines were just as easily missed by
professionals as they would have been by a group of students or
volunteers.  Often this was due to unrealistic or poorly thought-out
deadlines.  Brooks in _The Mythical Man-Month_ and McConnell in
_Software Project Survival Guide_ offer some insight into why this
happens and how to avoid it.

> if it's a minor package it can be dropped usually but if its something
> more important ..

A radical idea I've seen tossed about on d-devel and elsewhere is to
loosen the traditional thinking that each package has a maintainer
who (for the most part) is king regarding that package.  This often
dampens NMUs which could solve problems more quickly.  If a "community
effort" model were implemented where maintainers still "supervise"
packages but allow others to work on them, fixes might be integrated
more quickly.  THis is a very simplistic synopsis of the argument, but
it is a bit like the development model my colleagues and I worked out
which was rather effective for us.

> the past couple releases anyways the holdup has always been the boot
> floppies(from what I read anyways). can't install the system unless the
> installer works ..

I don't know how much debian (or SPI) collects every year in
donations; I suspect it's much less than it should be.  I see two
related possibilities here:

1) The DPL should appoint a committee (yuck) which sets up "rewards"
or "bounties" for important projects that need to be completed
yesterday.  A little green never hurts.

2) Debian should be more active in soliciting donations from people
who use debian.  Redhat users seem to think nothing of paying $30 for
a set of CDs; I bet many debian users could pitch in $5.  Corporations
running debian could probably afford more, or even earmark monies
towards specific projects (debian-desktop, etc.)
 
> debian does announce freeze dates though I think, I don't remember seeing
> anything for woody freezing yet, again the big project is boot floppies
> where they are doing a ground up re-design of them, hopefully it gets
> done in time, in previous releases they tried to do it, but ran out of
> time and had to resort to hacking the previous boot floppies up for the
> newer version(which I read is very difficult since the code is so old).

Ah, so in fact debian does have deadlines from time to time :-)

Yes, the debian release process can be frustrating, especially if you
haven't been around it before.  In my experience (my first install was
rex) debian does get better and better at dealing with issues each
release; however the size and scope of the project grows with each
release as well which means new problems are constantly uncovered.

debian actively tries to _solve_ these problems; very rarely are they
"swept under the rug".  This is cool.

To those of you who have a big problem with debian's release schedule
(or lack thereof), may I humbly make some suggections?

1) subscribe to d-devel, and get educated on what the issues are.

2) help out.  Test and file bugs.  Work on the installer or write code
if you are so inclined.  Write documentation (both software and
processes such as installation need documentation).

3) if you still aren't happy or can't (won't) do 1 and/or 2, please
try another distribution.  I don't mean this in a pejorative way; I'm
saying be a good capitalist and check out the competition.  If they're
better than debian, running debian is a waste of time.  If they're not
better, then perhaps 1 and 2 deserve more attention (or perhaps you
can donate a substantial sum to further some project).

As for me, I'll stick with debian.  It just works.  Could it be
better?  Sure, and I believe it will be.

-- 
Nathan Norman - Incanus Networking mailto:nnorman@incanus.net
  A young man wrote to Mozart and said:
  
  Q: "Herr Mozart, I am thinking of writing symphonies. Can you give me any
     suggestions as to how to get started?"
  A: "A symphony is a very complex musical form, perhaps you should begin with
     some simple lieder and work your way up to a symphony."
  Q: "But Herr Mozart, you were writing symphonies when you were 8 years old."
  A: "But I never asked anybody how."



Reply to: