Re: Some thoughts about problems within Debian
On Fri, 2002-01-04 at 17:29, Colin Walters wrote:
> > > The target audience is everyone who uses a computer.
> > OK, but that is not very clear a goal. What do we want with all those
> > people? Do we want them all to use Debian? If so, what should be done to
> > accomplish this goals?
> Yes, of course we want them all to use Debian; we claim to be the
> Universal Operating System.
In my opinion, that is not a sensible goal. Perhaps the "Universal
Operating System" should be thought of as a notion of availability, not
functionality. It is not the tool that solves all the problems that can
be solved using an operating system. Rather, it is an operating system
-- that solves many problems -- that is universally available.
Having an ideology of Free software does not mean you have to do
everything partially because there exists some moral obligation to make
everybody happy. You can do fewer things completely instead. Thus the
notion of target audience. Define your target audience, and it will be
easier for most people to see the intended direction. It is also easier
for an individual to determine whether or not he or she belongs to that
target audience. Thus, you let no one down.
In time, you may very well have different target audiences that,
combined, are the "everyone who uses a computer" audience.
In Debian, the technical representation of a target audience is in fact
already present in a certain form: task packages (or whatever the name
> What the exact steps should be to accomplish this goal aren't completely
> clear to me, and I doubt they are to anyone else either. We have come a
> long way, but there is a lot more to do. Just for example, take the
> thread about Euro support. Certainly, Europe is part of the Universe
> which we want to support. So, all of Debian needs to support the Euro.
> Well, how do we do that? For some programs this may require something
> as trivial as setting an environment variable, but for others it could
> be quite complex, requiring code/program design changes. So even the
> steps to getting to full Euro support aren't well defined.
A very good example. Unfortunately, it is an example of where Debian has
already failed to meet the obvious goal. The obvious goal would have
been to introduce Euro support in Debian in time for the official
introduction of the Euro coins and paper notes (1.1.2002) -- preferably
some time earlier so that people would have some time for the
Free software projects such as Debian often seem to suffer from
lateness. I think it is because of shortcomings in project management.
But project management has never been given a chance because Debian is
so huge. There is enough work to be done on purely technical issues --
but the requirements change faster than the technical issues can be
resolved. Thus, a huge backlog of work builds up, and we end up with the
current situation, without having the administrative routines properly
set up. Think about it, what would become of a large software company if
every time an employee quit (= MIA maintainer) there would be no routine
for reassigning his work?
> I think Debian is really defined not by its internal structure, or even
> things like the Social Contract or the DFSG, but by the people, who
> choose what to work on at will. We have to have some concrete
> definition (i.e. policy, the DFSG), of course, or things would
> disintegrate into chaos, but just creating structure doesn't make people
> work on filling it in.
> I guess what I am really trying to say is that the goal is so ambitious
> that we can only work on it in small chunks. So if one particular
> problem is bothering you (like Euro support), then start a subgroup
> within Debian to work on those programs, and create a list of things
> people can work on in that group. Then people who are interested in
> that general area can look at that list of specific tasks to do.
Yes. Debian is the sum of everyone who works on it. In fact, Debian is
more -- think about all the upstream authors and contributors to large
projects that produce some of the software in Debian.
However, a thing like Euro support is by itself a task as huge as the
Debian project. For things like that to work, more than a dormant
"do-it-if-you-can-be-bothered" group is needed. Someone (who has
volunteered for a position in Debian management) needs the authority to
say what and when, otherwise we will be stuck forever in endless
discussions such as this one.
Different Free software groups do project management differently. But
some discipline and clear goals do seem to benefit most of them. That is
not to say that you should prevent people from making just a one-time
small contribution. You should just try to make small groups of people
working in close co-operation, ie. split the things that need to be done
into chunks that can easily fit into the human mind. This is separate
from the "maintainer" notion and is not a replacement. Maybe this is
what QA is becoming?
PS. While rants and flames seem to be easily aroused at this time, let's
remember that a lot of hard work has gone into having anything at all to
even complain about. Woody is a very promising, if somewhat late
(feature-wise) Debian release. Many thanks go to the people involved.