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

Re: Debian as a social group and how to develop it better

Xavian-Anderson Macpherson => X-A+M, for further reference to my remarks.
\\ => designates end of remarks

On Friday 2002 November 29 13:10, Martin Schulze wrote:
> Andreas Schuldei wrote:
> > I have been reading books about group development and would like
> > to share the thoughts about appling this to debian.
> >
> > The book identified several different, interconnected "state
> > variables" for groups:
> >
> > empowering leadership
> > gift-oriented cooperation and work
> > enthusiasm for the group's goal
> > functional structures
> > holistic small groups
> > need-oriented advocacy
> > loving relationships
> > inspiring meetings
> >
> >
> > groups that rate high in all of these areas have been found to be
> > growing and thriving. They are much fitter then groups where one
> I believe that you missed a very important item in your list, probably
> because a group of volunteers is very different from a group of
> co-workers.
> Rules and regulations
> This is the most important thing in Debian that keeps the project from
> falling apart.  We, fortunately, have very precise rules about our
> business (social contract, DFSG, developers reference etc.) and a very
> good technical infrastructure (machines, katie, bts etc.).  Without
> these rules and our infrastructure, we would probably not be able to
> work with other 900 developers cooperatively.
> Also, anybody who wants to join as a Debian developer and contribute
> to the distribution needs to obey to these rules.  People mostly join
> because they believe in Free Software or they believe that Debian
> creates the best distribution of Free Software that they want to use
> and they want to support its development.  However, without our rules
> and infrastructure, things won't be clearly ordered and won't work in
> an automated and fast manner.
> > or more of these areas are underdeveloped. This approach can also
> > be used to identify and remove the bottleneck for further growth.
> It would've been helpful if you would have added in which areas Debian
> needs to improve and where our bottlenecks are in order to become
> "fitter".  (I've seen two issues in your mail.  If they are the only
> one, we're pretty good according to your books.)
> > the most effective way to grow (both in quality and quantity) is
> > usally to remove this bottleneck by finding ways/strategies to
> > develop this minimum factor.
> Another question would be whether unlimited growth is sane.

X-A+M:  Unlimited growth may not be possible, but a significant expansion must 
occur, unless Debian wants to remain as antisocial as FreeBSD.  If the 
enhancement of one's ego is the primary reason for running this OS, so that 
you can tell yourselves how superior you are for doing so, then I geuss you 
will become and remain a group of pendantic sobs.  I HOPE NOT!!  That's 
pendantic, as in "hanging on the fringes" of the greater social need for easy 
and aaptable computing environments.  As in having desktop icons for drive 
access, without having to figure it out on your own.
> > Empowering leadership
> >
> > It is obviously more important that leaders are good planners,
> > communicators and people persons then that they are excellent
> > technicans. The most distinct difference between good and less
> > good leaders is their ability to enable the others to do their
> > job well and give them authority in their area of responsibility.
> > They coach and councel the few who they are leading directly, not
> > micro-managing each and every one. It is important to notice
> > that there are several leaders in big groups on different levels,
> > not just one, above all others.
> I'd like to emphasize that leading and coaching are two different
> things and that Debian rather needs a coach than a leader.
> Remember Bruce Perens who once lead Debian and failed since the Debian
> people didn't want to follow his path, and finally he even quit the
> project.
> > Debian seemed to have elected mostly technical persons as their
> That's probably because Debian consists of mostly technically oriented
> people.
> The above probably refers to technicians in a company being lead by a
> manager who doesn't understand the technicians work.  This may be
> healthy for a company that needs to reach market saturation or
> something similar.  However, this cannot be applied to a volunteer
> group of techs.

X-A+M:  I would define "Empowering Leadership" as USER's RULE!!  What are the 
goals of Debian as they relate to the world community of computer users?  Are 
you wanting to remain just a step above the 'hobbyist notion' of Slackware, 
or do you truly want to reach as many users as is possible?  I have seen an 
article (on a Debian main page) on how to use Knoppix to provide an easier 
(GUI) method of installation for Debian.  My question is, why should I want 
to use Debian, if I am clearly being told by Debian the system is so hard to 
install.  I geuss this goes back to my question about distibution (in terms 
of population).

If someone else can build an easier method of installation (based on Debian no 
less), why is it not included in Debian as the standard, if the 'standard' is 
not to be interpreted as 'for  hobbyists only'?  I mean, even the install 
system of FreeBSD is decent, in that you can choose packages from the web, 
using just two floppies.  Why can't you develope an installation system that 
downloads a wonderful GUI from the web (or from cd), that provides all of the 
features of let's say YaST from SuSE.  I think the only requirement should be 
the prior creation of a swap partition, on which everything (package names 
and configuration files) can be temporarily loaded, until the installation is 
fully committed to.  This way, there would be no need for separate cd's for 
xfs (as with mine) or other system features that must be chosen before 
installation.  The only drivers you would need to provide would be network 
controllers.  Why not use KDE's control panel for installation, or even 
webmin (as is used by MSC Linux, and very nice!)  This way, whatever 
installation tool that was used first, will have been already taught to all 
new users from the beginnig, eliminating any possibility of subsequent 


Now, after having gone back over this, I realize that someone may say that the 
kernel is the limiting factor.  That is the kernel must be configured from 
the beginng to provide the features that are initially wanted.  Well this 
raises another point of consideration.  I remember from my days with the 
early Mac, that the operating system was on a single floppy, and that if you 
mounted a floppy with a more advanced kernel, the computer would "upgrade" to 
that kernel immediately.  This would necessitate that you constantly replace 
that floppy whenever needed (unless you copied that kernel to harddisk).  
This was a nuisance back then, but could be a solution now.  If it was 
possible for the archaic Mac to do this, WHY can't linux.  Even FreeBSD 
allows you to install features into the kernel without rebooting, whereas I 
think linux would require this.  Linux must be converted into a system that 
ABSOLUTELY NEVER needs to be rebooted for ANY reason at all.  Someone told me 
that HP has this feature in their kernel, the ability to upgrade kernels 
without rebooting.  This needs to be in linux as well.  NEVER REBOOT!!

Consolidation of (static) system resources into a cohesive (installation) unit 
is the best way to avoid unforeseen problems and consequently obstacles to 
new users.  The ONE program I CANNOT do withour is MC.  But because I had to 
tell FreeBSD where the program was (because I couldn't simply type mc on the 
commandline and get mc), that was one of the first reasons why I stopped 
using FreeBSD.  With MC, I can do just about anything I need to do.  So if 
everything else fails, as long as I can get back to MC, I feel safe.  
Whatever is decided upon for this purpose must always be kept as STABLE, and 
only available as such, except for the maintainers.  Any changes that are 
made must be made by them from suggestions by general users.  This package 
needed for the operation of this package must be included in the package, NO 
EXCEPTIONS!  Developers of climbing equipment don't send users into the field 
to test safety equipment; this is restricted to the lab.  Whatever the 
standards are for the developement of security patches in Debian, needs to be 
the same standard for this installation package.  Because this installation 
package needs to thought of as a security issue.  It would be the one thing 
that controls everything else from that time on.

That same sort of security net needs to be provided to everyone.  Especially 
newbees!!  Give us one resource (like MS' w2k recovery conole, that's truly 
intuitive), that no matter how bad things get, we can always count on 
whatever "it" is, to recover from all or most of any errors we may encounter, 
and I PROMISE you, Debian would take over the world.


In fact, MC (with zsh) would be a perfect candidate for this purpose.  Make MC 
the universal solution to all installation and maitainence issues, (with the 
addition of modules to MC), and you will have solved anything that anyone 
could want.  (And put this program on the /BOOT partition, so that no matter 
what happens, it's ALWAYS there.)  Either that, or provide a mixture of 
linuxcong and mc (webmin would be better) and all your problems (and mine 
too) would be solved!!

Let me define "better", from my mention of webmin above.  I mean better as in 
easier on the eye in terms of intuitiveness.  There have been studies as to 
what features humans universally perceive as beautiful.  These features were 
then quantified so as to give a generalized image of beauty (male and 
female).  The resulting image(s) transected most if not all racial and 
culteral boudaries.  There are features that humans universally tend (by 
intuition) to use to determine if or how much a prospective mate is fertile 
(eg, by body contours and proportionality; to facilitate the global mandate 
for survival of the species).  This would correlate to user friendliness and 
ergonomic placement.  You could say that such an image of beauty would be 
familiar to mostly every person (ie, computer user in my analogy).  It is 
this sort of familiarity that must be established in Debian, ie linux.  
Debian has the power of leverage, if for no other reason than the number of 
packages available, to influence in a decisive manner, every other 
distribution of linux.  The rest have already copied APT-GET.  To whom much 
is given, much is required.


I have used ALIEN (love it) to install packages from SuSE and others, but the 
best thing would be to have the Debian package database copied as the RPM 
database, using the RedHat package "rpmdb-redhat" as the reference, so that 
any packages installed are recognized by both.  Granted, some of what I am 
talking about may already exist.  But if so, it only serves to prove my point 
of not being intuitve.  The one thing that MUST be agreed to by EVERY 
distribution of linux, is that ALL package names and locations MUST remain as 
ORIGINALLY INTENDED by the package maintainers, WITHOUT EXCEPTION.  This is 
the one major fault I have had with SuSE.  It is excellent in it's user 
friendliness, but the addition of other packages from foreign sources cause 
nothing but trouble.  And now that SuSE has joined the UnitedLinux 
consortium, I really don't think things will get better.  The bottom is that 
SuSE wants to be MS!  Such aspirations are hostile to linux.

If Debian can clean up it's act, regarding it's installation, Debian would 
become the premier linux system.  That is something that could not be 
debated.  But Debian MUST be a system FOR the people, by the people, or else 
you must be honest and say forthrightly, that is not your desire or 
intention.  Because right now, you are very successful at discouraging new 
linux users.  And by means of this discouragement, you are directing them 
everywhere else but here.  And if you (GNU) want to maintain the availablilty 
of FREE software, the software MUST be FREE in terms of it's ease of 
installation.  Otherwise, you are deluding yourselves, becuase no one else 


I chose SuSE over RedHat two years ago because of value by price.  SuSE 
provided 6 cds to RedHat's 3, for a price half as much.  That same day, I 
came to the store to buy Debian.  I was convinced based on what I read, that 
Debian was better.  But when I was confronted with the reality of pricing, 
Debian was the most expensive.  Debian only had one cd to RedHat's 3 and 
SuSE's 6.  I went with SuSE.  The price of Debian and SuSE were the same, 
$49.99.  Debian had an encyclopedia for a manual, SuSE had a dictionary (by 
comparison).  I would have killed myself if I had gone with Debian.  The only 
reason why I am using Debian now, is because I have used linux enough to know 
how to fix some things when things go bad.  Which for me is often, because I 
am always doing things no sane person should.  This is where sanity is 
defined as not mixing packages from other distributions, because no one 
distribution has it all.  I wouldn't recommend Debian to anyone who is not as 
crazy as I am.  I have enough respect and desire for the prosperity of Linux, 
not to do so.  Any new user attempting to start liunx for the first time 
after coming from Windows, would be absolutely angry with Debian.  And the 
fact is, this FACT is NOT necessary.


> > DPL, with the result that their success to innovate and
> > reach their goals was limited. In other areas delegates of the
> I wonder if you understood how innovations happen in Free Software
> projects such as Debian.  It's not like within a company where the
> leader says "We are doing foo" and the staff has to follow (or won't
> be happy about their next paycheck.
> If our leader announces that "we do foo" and the developers disagree
> and even want to do bar, he's in a lost position since he cannot
> enforce foo.  All he can do is get something like consensous for foo
> and persuade those who still would like to do bar and if he's
> successful, announce that we do foo.

X-A+M:  I geuss this is a problem, if "the needs of the greater public" are 
not the concerns of the Debian group!  If you were inclined to think of the 
public as your leader, you would have no legitimate means of arguing with the 
> > DPL try to let no one interfere with their area of competence,
> > while they, as leaders themself, would be wise to find others
> > interested in their center of competence and in turn educate,
> > train and empower them. (here the keyring management comes to my
> > mind)

X-A+M:  Again, make 'the needs of the public' your leader, and this won't be 
an issue!!
> It is not always wise to split the work among more people.  More
> people imply more coordination, better tools, stricter rules and
> perhaps a high level of trust (here the keyring management comes to
> my mind).
> > During the NM process the understanding of the social contract
> > and the GNU part of Debian is checked for. it is hard, but not
> > impossible to check for the actual enthusiasm of people. What
> > debian has not managed well in my point of view is keep
> > enthusiasm alive or newly create it. You can see that in the
> So what's the path Debian should go?
> We encourage people to report bugs.
> We encourage people to send patches.
> We encourage people to integrate (more) software.
> We encourage people to help with documentation.
> We encourage people to develop software.
> What's missing?
> > large number of people who became Debain developers and dropped
> > out over time. some Debian longtimers are beginning to show signs
> > of burn-out. the technical commitee is one example. If someone
> > quits and says: "i have not enough time for debian anymore" he
> > actually is stating that the priority which he gives debian is no
> > longer high enough and that he is not excited about it anymore.

X-A+M:  I personally think the greatest defense against burn-out would be the 
positive response of the general public by a broader acceptance of Debian.  
There are only a limited number of geeks, nerds, and hobbyists that can 
support a system for people of 'like-mind'!  There just aren't enough of them 
in the world to dictate that "knowing everything about all there is to know" 
is better than being able to turn on your computer and use it for what you 
want when you want it!!  I have no problem with control, I love it.  But 
there are some things I don't want to know.  It would be kind of like having 
to do the g in a new car before you get to drive it.  You can just imagine 
how far that would get a company.  I don't care how well made the car was, it 
could be a Lamborghini, and I would not want to have to 'wire it' before 
driving it!
> I don't have the feeling that most people who are in the CTTE are
> burnt out but they can't afford as much time as they could five years
> ago.
> > Debian has great tools, a great infrastucture, the policy, which
> > structures the packages, the constitution, which determines the
> > democratic processes, and the cabal, which determines what
> > happens. (c: Recent discussions about gentoo seem to indicate
> > that younger distros with less restictions are more attractive to
> > some people then debian is. to me it appears to be really
> I believe that Gentoo and Rocklinux are for people who have different
> goals than Debian.  For those, Debian is simply the wrong distribution
> and they only used it because there was nothing suited better.  Now
> there is and they're happy.
> I'm stopping now.
> Regards,
> 	Joey

Reply to: