\documentclass{article}

\usepackage[latin1]{inputenc}
\usepackage{xspace}

\usepackage{debian-links}
%\usepackage{debian-freebeer}
\usepackage{fancybox}
\usepackage{alltt}

\title{Knowledge, Power and free Beer}
%\subtitle{... oder warum Medinzinsoftware so ist wie sie ist}

\author{Andreas Tille}

\date{June 16, 2007}

\newcommand\meinbs{\char 92 }

\begin{document}

\maketitle
\tableofcontents

\section{Free Beer}

\subsection{Hard and soft Ware}

\subsubsection{Hard Ware}
In the field of informatics we differentiate between hardware and
software. What does these terms mean?

First lets think about the term ware in general. A key feature of a
ware - which is something which is selled by a vendor - is its weight.
Very often the price of the ware is proportional to this weight. Both
parties - vendor and vendee are able to do certain things like biting on
the thing or throbbing at it or whatever.

All these methods might be useful to check the quality and
thus the value of the ware to buy - or just for fun.  In any case
the buyer is able to check whether the ware fits his
expectations before handing out the money while the seller is
able to prove that the features of the thing justifies the price.

The term computer hardware perfectly fits this features. It comprises
all of the physical parts of a computer, as distinguished from the
software that provides instructions for the hardware to accomplish
tasks.

Once the vendee has bought the ware he deserves certain rights on it.
He is now the owner of this ware (provided he has paid the full price
he and the vendor agreed to).  The person who owns something has the
pristine right to alter, change or disassemble the thing.  If the ware
shows some hidden problems, has some faults or gets broken in a
certain time the vendor has to guarantee for the promised features.
Finally the owner has the right to sale again the ware that was bought
some time before.

As a rule duplication of a ware is not worth the effort because the
professional manufacturer mostly can do it much cheaper.

\subsubsection{Soft Ware}

The features of software are quite different.  In fact the question
comes up whether it really deserves the term ``ware''.  The main
feature of weight is missing completely because defining a weight for
software makes absolutely no sense.  Moreover the person who is about
to buy a certain software is barely able to check it completely as it
is possible with any ware.  Even if there are some vending models to
use a software for a limited time before buying it.  But these days
software is this complex that the verification and comparison of
competing products is sheer impossible (compared to find out which
salesman trades fresh fruits at the market place).

But even if the buyer payed the price for the software the relation to
the thing which was payed is quite different than for a piece of
hardware which was discussed above.  This becomes clear when somebody
would take the time to read the strange text which is always displayed
of the installation routine of proprietary software.  There must be a
reason why this hard to understand text which sounds very juridical is
bothering the user who wants just to install a piece of software on
the computer.  But most users have learned that it works perfectly to
just press the \texttt{``Next''}-button while ignoring the text.

This behaviour prevents users from recognising that they do not really
own what they payed money for, they just got a license to use it,
sometimes applying some additional restrictions like that the software
can only be used on this specific computer and will only work if you
send in some data of your computer to get a license key.  In case of a
piece of hardware -- say a chair -- this would mean you pay for the chair
but you are only allowed to use it in a certain room next to a
specific table and only at the gable end; for the other side a
different chair with a license to use at this place is needed.

Moreover the license of proprietary software does not allow to change
anything.  Sticking to the chair-example, this would mean the user is not
allowed to paint it with a different colour.  Last but not least
selling the software is often not allowed.  The reason is obvious: The
user just does not own the software - the money the user spend on it
was for a license to use but not to sell.

Besides this completely different relation between the owner of a
piece of hardware and the license owner of some software there is a
further difference: In principle it would not be hard to duplicate
software.  That's why this is not only strictly forbidden but also
prevented by technical means.

\subsubsection{Fresh fish}

% http://www.comedix.de/lexikon/db/verleihn.php
%%% Verleihnix
%%% Orginal heißt dieser Charakter Ordralphabétix
%%% Englisch: Unhygienix (UK), Epidemix (US)
Around 50\,BC at the market place of a small village in northwest
Armorica the fishmonger {\em Unhygienix} is infamous for his
rotting products.  This often causes trouble with
{\em Fulliautomatix} and other inhabitants of the village. This
means selling bad smelling fish at the market place is nearly
impossible.  Each customer would immediately notice that it is to
old.

But what about software?

The customer has no sign about how old the code of proprietary
software might be.  It has no odour or other signs of age which are
easily to recognise.  Moreover software usually consist of several
parts which are developed at different times.  It frequently happens
that the major parts are quite old and only new features and user
interface is really fresh.

But what is the problem here?

If somebody pays for a license a certain amount of money only a
fraction is payed for really new code.  A larger part of the money is
spend for code which yet was paid in former versions.  But former
versions should not cost money.  Users just got the license to use
this code for years.  But what are the stinky bits in the whole
software bundle?

Lets recapitulate: Users are perfectly willing to pay for the work of
the programmers.  It is a fair deal to pay for a service like software
development.  But the deal is only fair in the case that the user is
able to verify how much time was spent to produce this piece of
software which was not payed before.

\subsection{Patent recipes}

\subsubsection{The progress}


The progress bar is something which keeps impatient users happy.  It
is the way of programs to inform users that some business is going on
and give hints how long this business will last.  It is no rocket
science.  If you see a 100\,m sprint you can clearly see start and goal
and the runners move like a progress bar from the beginning to the
end.  So a progress bar can also be viewed as a analogue to real live
procedures which are just visualised at the computer monitor.  It is
hard to tell what the technical invention in the progress bar might be
so it hardly can be patented.

But there is \texttt{Patent No.~EP0394160; Applicant(s): IBM (US)}
which exactly describes the progress bar.

The next example for a so called ``technical invention'' might be
another illustration of the issue: It seems reasonable to store
different versions of a document in that way that you include the date
in the filename.  Provided that this date-tagged file archive is
stored on a network drive which might be accessible to colleagues than
this ``technique'' infringes \texttt{Patent No.~DE10108564} which is
just valid in Europe.

So what is the issue in declaring common sense as technical invention?

Big software companies acquire a large portfolio of patents -
reasonable but mostly such strange things as mentioned above. In case
they want to use a technique which is patented by another big company
the deal is as follows: Allow me to use your technique and I allow you
to use my technique for free.  So a large amount of patents is kind of
an insurance not to become sued by a competitor.  This sounds very
reasonable from the point of big software companies and is the driving
force behind them to register as many as possible patents.

But what about smaller software companies that do not have this kind
of insurance.  Or even worse what happens to the Free Software
programmer who has not even the slightest idea that his idea he had
from common sense is patented by someone else?

\subsubsection{EU: Council versus Parliament}

This talk is about Free Software.  It is not about politics.  But
unfortunately the topic of software patents in Europe is a very
political topic.  Unfortunately not many people do care about this.
So it remains mostly not regarded if representatives of national
parliaments vote for different things than the national parliament has
instructed them to vote in the Council of Europe.  Less people notice
if the rules of democracy are broken for the sake of big software
companies.  But this would be material for one or two separate talks.

So here is some homework for the interested reader.  Just ask your
favourite search engine for the following questions:

\begin{enumerate}
  \item What is the difference between the Council and the Parliament
    of Europe?
  \item Which panel brought up the issue of software patents and what
    is the suspicion who is behind this issue?
  \item What do we owe the representatives from Poland in the Council
    of Europe?
  \item If you are not from Poland what role played the
    representatives from your own country in the Council?
  \item Which panel has the chance to stop the efforts in the wrong
    direction.
\end{enumerate}

If you are able to answer all these questions you can go straight to
your representative which you voted for the European Parliament and
ask him to vote against software patents.

\subsection{Free of charge}

  \subsubsection{Soft ware is not for sale}

In the beginning it was stated that software can hardly be regarded as
a ware.  This has the consequence that it is hard to sell it for a
fair price.  There are many people who get this idea but came to
different conclusions.  Common to these conclusion is the fact that
they give away their software at no cost.  But what are the
differences.

\begin{description}
  \item[{\em Download here}]
    The user can download the ready to run executable. For most users
    this is fine.  They want a program which just does a certain job
    and do not want to pay money.  This is kind of ``free beer'' -
    software.  You go to the restaurant to get drunk without spending
    money - free beer works perfectly for this purpose.

    But is this really freedom?  Does this program do what the user
    expects it to do?  Isn't it just a Trojan horse which spies out
    user data?  Can the user be really sure that it is not harmful?

    It just happened that people mixed something into the free beer
    \dots

  \item[{\em Do not change}]
    The author of a piece of software allows the user to have a look
    at the code.  There might be several conditions under which the
    user is allowed to look at the code but assume here that having a
    view on the code and using the program is free of charge.

    This at least might ensure the user that the code does not contain
    any harmful bits which is not the case in the {\em free beer}
    flavour of free of charge software.

  \item[{\em Free Software}]
    Free Software is according to its definition not only free of
    charge, it also allows to use the code in complete freedom which
    means the user is allowed to change the code for his own purpose
    and can even give away the changed version.
\end{description}

If there is such a lot of software free of charge the world must be a
better place.  So a normal user might wonder: Where's the catch?


\subsubsection{The catch: Free $\ne$ free of charge}

The term ``Free Software'' does not say anything about the money a user
has to spend to solve a certain task with this very peace of
software.  The crucial thing is just that the code of the software can
be used without any restrictions by anybody.  The fact that the code
can be used by anybody free of charge does not implicitely mean that
it is really useful to solve a certain task immediately.  Any piece of
software needs adaptations, maintainance and service.  This does not
come for free.

The {\em free} in ``Free Software'' is more in the sense as ``free
speech'' instead of free of charge.  Free speech requires time to
write out.  It needs competence about the topic and sometimes free
speech requires courage.  All these are values which can not be
countervailed with money -- but they are certainly worth anything.  So
there is either some cost of time for the user in person or the user
has to pay anybody to do the job.

The exact definition of Free Software is given in the \DFSG.

\section{Knowledge}

\subsection{Alternatives}

 \subsubsection{WikiPedia}

If there is anything in the world which is easily comparable to the
fascination of Free Software by sharing many ideas in common than it
is free knowledge.  The goal to collect the whole knowledge of mankind
for the whole mankind for free is the goal of
\link{http://wikipedia.org}{WikiPedia} for short.

If you are an expert in any field or think you know something special
just test the quality of this community effort driven encyclopedia: Go
to \link{http://wikipedia.org}{wikipedia.org} and look up this
special field.  There are two basic options:
\begin{enumerate}
  \item The contents is well written and is right according to your
    knowledge. Just be happy that you found a new information source
    and continue checking the contents of other fields if they
    represent your own knowledge.
  \item The contents is not well written, is not precise or even
    completely wrong - you just are not happy with the contents. There
    might be people who just surf away and use other sources of
    information while ignoring this WikiPedia rubbish.

    But once you proceeded through this paper up to this point you hopefully
    behave differently.  You learned something.  You just recognised the fact that
    nothing comes for free.  Take a little bit of your spare time and
    bring your knowledge in precise and well formed sentences and just
    make WikiPedia better.  Your fellow WikiPedia writers will be
    happy about this and other readers will be even more happy from
    now on once they stumble over the page which was not satisfying
    before your editing.
\end{enumerate}

 \subsubsection{WikiPedia is taken honestly}

%%Content analysis details:   (12.1 points, 5.0 required)
%%
%% pts rule name              description
%%---- ---------------------- --------------------------------------------------
%%-2.0 USER_IN_WHITELIST_TO   User is listed in 'whitelist_to'
%%-3.0 ALL_TRUSTED            Did not pass through any untrusted hosts
%% 0.1 SATIS_GUAR             BODY: Mail guarantees satisfaction
%%  17 BAYES_99               BODY: Bayesian spam probability is 99 to 100%
%%                            [score: 1.0000]

\begin{figure}
\begin{center}
%\thicklines
\ovalbox{
\hbox to 0.7\textwidth{\hss
\parbox{0.6\textwidth}{
\begin{alltt}
~ \\
{\bf Subject: {\em Wikipedia} knows the shit} \\
Date: Sun, 27 Mar 2005 16:18:36 -0200 \\
From: Christian <VPYLGHQ@interacti.net> \\
To: aids-std@rki.de \\
\\
Newest penis enlargement system: \\
 - Increases the penis length and girth \\
 - Medically Proven (source: {\em wikipedia}) \\
 - No Surgery \\
 - Permanent Results \\
 - Proven Traction Method \\
 - 100\% Satisfaction Guaranteed \\
\\
Find more info here: \dots \\
\end{alltt}
}
\hss}
}
\end{center}
\caption{Spam mail referring WikiPedia}\label{figure:spam}
\end{figure}

Figure \ref{figure:spam} shows a quite interesting spam mail.  Well,
not the contents itself is interesting, but the fact that spammers do
refer WikiPedia to ``prove'' that their rubbish is worth anything.
This is kind of: If your enemy takes you honest you must be somebody.

On the other hand it reveals the most common biased opinion about
WikiPedia: If {\em anybody} can write and change articles how can be
assured that nobody writes something which is wrong on purpose.  (This
effect is called ``vandalism'' in the WikiPedia slang.)  The answer
is: Many people are watching.

There is a log of every single change of each single page.  Reverting
some change is done quite quickly and thus the effort of vandalism is
high compared to the effort which would be needed to keep WikiPedia
clean.

Assume the dirty spammer which sended the spam mail which is displayed
in figure \ref{figure:spam} had edited WikiPedia
first before sending out the strange mail.  Even if this person would
have done this -- it is not in WikiPedia any more.  This is one
example that the auditing process of the WikiPedia activists is
working effectively and even if there is vandalism it does not seem to
have a major influenze on the whole collection of free knowledge.

Last but not least there is some solution for the boring spam problem:
In figure \ref{figure:spamassassin} the output of this spam filter
showes that it catched the beast -- a nice example for a useful peace
of Free Software.

\begin{figure}
\begin{center}
%\thicklines
\ovalbox{
\hbox to 0.95\textwidth{\hss
\parbox{0.85\textwidth}{
~ \\
\texttt{\bf SpamAssassin: 17.1 points, 5.0 required}
\small\tt
\begin{tabular}{rll}
pts & rule name  & description \\ \hline
0.1 & SATIS\_{}GUAR & BODY: Mail guarantees satisfaction \\
 17 & BAYES\_{}99   & BODY: Bayesian spam probability is 99 to 100\% \\
\end{tabular}
~ \\
}
\hss}
}
\end{center}
\caption{SpamAssassin handled spam mail}\label{figure:spamassassin}
\end{figure}

\subsection{Free Software in general}

  \subsubsection{\label{UserBase}As You Like it}

What is the basic need of users of any piece of software? What do
users really want to solve a certain problem?

A steady and continuous functionality over long period is the main
demand of software users.  The program has to be safe against failures
to not interrupt the work.  It has to have a certain set of functions
which are necessary to solve a certain task and it has to provide this
functionality for a long period in which the task in question has to
be solved.  Any upgrades of the software do require new learning which
just takes time which is cut of from the time span which is reserved
for the main task of the user.  It is a common missconception that
users always want the latest and greatest version of any software
project.  A further missconception is to overload programs with more
and more functions: The more functions a program has the harder ist
will be for the user to recognise which functions are really helpful
to solve the task in question.

If these requirements are fullfilled the user does not really matter
about the concrete program which is finally choosen to solve a certain
task.  A user simply wants to solve a task and does not want to bother
about the philosophy of the tool that is used for it.  Just ask your
local carpenter whether he cares about the wood which was used to make
a handle for the sledge he is using.  Most carpenters will not care
whether this piece of wood comes from a forest which should be saved
or from anywhere else.  So users just want a functionality.

In the case of software functionality is not just the program.  It is
more than that: it is the program and the service which is provided
for it. Users have to decide how much money they want to spend for the
given task - usually there is some budget for any task. So if there is
a given budget and the program itself comes for free there is more
money left for the service.  Usually service is quite expensive these
days so it makes sense to spend more on this side of the calcualtion.

So far to the advantages from the users point of view.  But what about
the producer of a piece of software?  What would a software company
loose if they would release their code which proprietary software
companies keep as a secret.  There are several reasons which speak for
releasing the code as Free Software.  It is just the problem that
everybody thinks it is impossible because selling proprietary binary
code is kind of normal these days.  But it is really the most
effective form to to make money in software business?

At first there are software companies which make all their money based
on Free Software like \redhat.  At second there are companies that see
their future in releasing at least parts of their code as Free
Software.  Well known examples are \novell or even \ibm.  The
rationale behind this is quite simple:  The proprietary software
market is currently characterised by a hard competition.  If a company
recognises that it is impossible to beat the competitor it tries to
drift to a new field.  In several cases this leaded to a release of
code as Free Software (Mozilla, OpenOffice, Interbase, MaxDB,
E-Directory, \dots) and providing service for this software (see above
for the relation between functionality service and program code).

As a matter of fact providing services for Free Software is simply
cheaper than if the code is not available.  It is quite common that
there are always some users of a certain piece of software that are keen on trying to
solve a problem they have themselves.  In case they have understand
who Free Software works -- sharing ideas, work together, join forces
-- they provide solutions or fix bugs in the software for free.  They
just have learned that if they provide their work to the authors the
next version will contain their code and will make their own work
easier because the program has an enhanced functionality.  And this is
the real clue: The authors of a piece of Free Software get freelancer
for no extra cost.  The driving force behind these freelancers who
volunteer to provide enhancements is that they need a certain
functionality to solve a task.  They want to spend this time only once
and thus they submit their work to the authors.  If the authors are
part of a company which provide services for this software in this way
they got some freelancers at no cost.  Compared to external
freelancers for proprietary code which requires hiring educated
programmers for a lot of money, signing discretion contracts while
being unsure whether they do the same for a competitor etc. the Free
Software approach has an extraordinary financial advantage for the
authors of the code.

The crux hereby is that you need a wide user base because only a small
amount of users will really provide reasonable input.  On the one hand
this is the case just because in most cases users will not be gifted
programmers on the other hand users might simply have not understand
the principle of Free Software.  They regard software as a ready
product which just has to be accepted with all flaws it might have and
with all problems it might cause when trying to do the tasks a user
want to do.  To solve the latter problem texts like this one are
written.

A wide user base to attract gifted programmers is normally no problem
in case of pieces general software which is needed by every user.  So
finding programmers for an operating system kernel (like Linux), a web
browser (like Mozilla), an office suite (like OpenOffice) etc. works
perfectly.  The problem changes in case of specialised software which
is naturally used by only specific user like people working in certain
professions.

%\pgfdeclareimage[interpolate=true,width=\paperwidth,height=\paperheight]{ms-linux}%
%                {ms-against-linux-response-advertisement.jpg}

\subsubsection{Get the facts}

When the main global player in software business recognised that there
is something else which is penetrating the software market.  After a
certain time they started a campaign title: ``Get the facts''.  They
paid for doing studies which should prove that proprietary software is
better than Free Software.  No matter what the result of these studies
might have been it shows one important thing:  They would not pay for
studies against somebody who is not taken honest.

BTW, it turned out that a certain amount of the facts was quite
questionable\dots

%There are some main differences between the 
% Unterschiedliche Sichtweise von Fehlern 
% Marketing $\rightarrow$\xspace Teamwork

\section{Power}

\subsection{Distribution}

\subsubsection{Linux from scratch}

The usual way a user obtains a software bundle is when buying a piece
of hardware.  This makes perfectly sense because computer hardware is
completely useless without software.  Usually a computer comes with a
proprietary operating system and some additional proprietary programs
to solve simple office tasks.

Assumed a user have heard about the existence of Free Software and
decided to try it on the new computer.  The reasons might vary from
the intent to save money (and buying a computer without any software)
to just being not happy with the software which would usually come bundled
with a computer.

The first thing which is needed is a Kernel of an operating system.
There are more than one option but Linux is the most popular
choice.  In principle the way to obtain a piece of Free Software is to
go to the web page where the source is provided, find a compiler which
is able to compile this source and install the program that was
builded on the machine.  It will take a certain amount of time to
install the Linux Kernel and some basic tools which make the
``GNU/Linux'' operating system.  This is described in the 
\link{http://www.linuxfromscratch.org/lfs/}{Linux from scratch}\xspace
project.

The modular nature of Free Software implicitly means that at this
point the user does not have a shiny colourful desktop but just a
command line interface.  The next step would be to get the source and
compile the X Window System which is not really a simple task for a
beginner.  If our user managed the work up to this point he reaches a
crossroad:  The religion of the desktop environment.  The main roads
are marked with the signposts ``Gnome'' and ``KDE'' but these are not
the only ones.  The user fails to find a reasonable decision and flips
a coin which road to follow.  From this point on the screen of the
computer looks nice -- but there is no program which enables the user
to do productive work.

Our brave user now has heard about two shiny new projects: Firefox and
Thunderbird. He finally manages to get these programs installed on the
computer and can do the first tasks: Browsing the web and managing
e-mails.  Even if this owner of the new computer started very
optimistic with the intent to install Free Software recognises that it
was quite a large amount of work to reach this step.

Once it comes to more complex user applications like a comfortable
office suite like Openoffice.org, a professional type setting system
like \LaTeX\xspace or a feature rich image manipulating program like
Gimp the user is really tired.

If it finally comes to some server software like a web server and a
database server the owner of the computer is completely tired.  So
many decisions to choose from, so many web sites to go - what is the
sense to spend so much time?

\subsubsection{Distributions on the market}

The answer is: Nobody really wants to spend so much time just to
exercise the possibility of Free Software: Do it yourself from the
source to the installation for every single piece of software.  It is
not only time consuming it is even error prone to build a complete
system.  It just should be done by experts.

A complete system containing all the bits of Free Software mentioned
above and much more is called {\em distribution}.  It is builded by experts
that know the code, use sophisticated tools to build the software and
prepare it to be installed quickly onto users machine.  This
preparation is done in so called {\em packages} that contain compiled
code of the Free Software projects and some preconfiguration to work
together with other components of the whole system.  Building a
distribution from Free Software is kind of a service.  Users may
decide whether they want to spend days for doing it all on their own
or just pay a certain amount of money to buy a distribution.  Most
users decide it is worth spending some money for a distribution.

But this uncovers some misunderstanding in the relation of Free
Software and money:  While (at least most) bits of the distribution
are free of charge the whole thing costs a certain amount of money.
This is one flavour of the ``a thing is more than the sum of its
parts'' principle.  The price which is payed is for the service inside
and users pay what they regard worth the money.  They just have to sum
up the amount of money they could gain with productive work in the
same time when doing all things on their own.  As a side note it should
be mentioned that there are also distributions available for no charge
but this is explained in detail below.

It has made clear now why users go to a software store and buy a
distribution of Free Software.  They buy a certain box which has
printed in bold letters: ``\textbf{XYZ} Linux version \textbf{A.B}''
where \textbf{XYZ} stands for the company that compiled the
distribution and \textbf{A.B} for the version number the distributor
has chosen to mark this release state of the whole bundle of
software.  New users tend to ignore the fact that this is by no means
the version number of Linux which is just the kernel of the operating
system and currently has a version number below 3.  So if somebody is
telling you that he is using ``Linux version 8.x or 9.z'' he has
apparently a misconception about what Linux is and what he really
bought in the software store.  To express this more clearly Free
Software enthusiasts speak from ``GNU/Linux'' because the complete
system is more than just the Linux kernel but the many other Free
Software tools around which finally make an operating system.

But once we come to colourful boxes with bold letters on it we are
back to the commercial world of market place with pressure of
competition and gaining market share.  It is a known fact that selling
GNU/Linux distribution is a normal business where all the rules apply
that are valid for vendors of proprietary software.  If a distribution
company is not able to make profit it will not be able to survive.
The good news is that there is a certain number of companies that
survive perfectly since several years.  Why is this good news?  It
just proves that it is perfectly doable to run a business on top of
Free Software - which is free of charge in its pieces.

But how to compete with other distributors that are just compiling the
same basic set of Free Software applications.  There are several
distinctions between distributors.  One is the local aspect.  So if a
user from Northern America speaks about his new Linux 5 (see above)
he is talking about his copy of RedHat Linux while a user in Germany
has not really a more recent version of the Linux Kernel if he is
talking about his new Linux 10 because he is referring to his copy
of SuSE Linux which has just a different version number scheme.
People in France prefer Mandrake Linux and Turbo Linux is the main
player in Asia -- except for China where RedFlag Linux seems to be
used mostly.

To make the jungle of distributions even more impenetrably there are
distributions that target more to the general desktop user with the
latest office applications and providing interfaces to proprietary
applications which normally run on other operating systems using
emulators (like Wine or DosEmu).  Users are free to pick the
distribution that fits their needs best.  They have a real choice
which they do not really have in the case of proprietary systems
because there is basically one for a given hardware platform.

Users have a chance to identify themselves with their distribution of
choice and who ever have heard over a flaming discussion between users
of different distributions knows what competition means.

\subsubsection{The missing link}

As it was explained in section \ref{UserBase} Free Software works most
effective if there is a wide user base.  That's why also distributors
decide to include these Free Software projects that are potentially
used by many users.  From a distributors point of view this makes
perfectly sense because serving a mass market implies the distribution
will be bought potentially by a high number of people which keeps the
business running.  It would be wasteful for the distributor to employ
specialists for preparing specialised software of certain fields to
sparse specialists the work to find out which software fits their
needs and how to install this software.

For a moment this sounds fair enough to build a solid base of often
needed applications and leave the remaining things for those people
who really need it.  But the problem is that because applications of
special fields do not have a wide user base and thus did not developed
so far that failure prove installation methods and easy updating
mechanisms are well developed which enable people who are no computer
experts to easily proceed with the task to install this software on
their computer.  The problem becomes worse if the applications require
a complex database or web server setup.

One solution would be that small service providing companies take over
this job.  It is quite common that the authors of a certain piece of
specialised Free Software provide their code for free and sell the
service of installation, adaptation, documentation and maintenance. 

The other solution is that a group of experts builds a complete
distribution by just moving the Free Software development principle to
distribution level.  In fact this is done and it got the name Debian.

\subsection{Debian - a distribution builded by a community}

\subsubsection{Maintainers are experts}

In contrast to the GNU/Linux distributions mentioned above Debian is
not a company but an organisation.  The goal of this organisation is
not to sell the distribution.  Debian sells nothing.  It just produces
a free distribution.  The people behind Debian who are united in this
organisation have just a common goal: To build the best operating
system they can afford.  While this goal sounds simple the motivation
for doing this seems to be missing.  While commercial distributors try
to earn money Debian people seem to work just for fun.

Well, sometimes it is real fun but there are stronger reasons for the
Debian people to work on their common goal: They just get what they
need for their own work.  They know that they can trust their system
and users appreciate their work.  The best appreciation users could
express is by just providing informations about problems they see,
suggesting solutions for these problems or even sending in code which
just has to be applied to make Debian better.

But how does this particularly work?  The developers of Debian are
volunteers who have a special interest in at least one special piece
of Free Software.  A developer becomes a so called maintainer of a
package by just doing the usual work all distributors are doing:
Obtain the source of the software from the download area of the
software, compile the software and build packages which are easy to
install.  This is so far no difference to other distributors (note:
the packaging format might be different but this is really a technical
matter).  The difference here is that this maintainer becomes on the
one hand the first user of this package -- so he is interested
personally -- and he is the first person to ask if users would run
into trouble with this package.  It becomes clear that the maintainer
is kind of the missing link between the authors of a Free Software
project and the final user.

The maintainer is probably one of the most qualified person for this
special application inside Debian - and the users know his name.  For
complex applications usually there is even a group of maintainers who
care for larger projects.  But each of them knows about the problems
that the application might cause in the installation process and tries
to solve these with inside scripts and configuration files that reduce
the work for the final user to a minimum.

\subsubsection{Special applications}

But what about the special applications issue which was not solved by
commercial distributors who are only able to provide a quite general
system?  This problem is solved inside Debian because also Debian
maintainers work in special fields and need special applications.  So
if a maintainer works in the field of medicine he will care for the
medical applications he needs.  Teachers can become Debian maintainers
and will probably care for applications that can be used in their
schools for education etc.  In Debian this principle is called {\em
  ``DoOcracy''} = the doer decides.

So it turned out that Debian became the largest collection of ready to
run Free Software in the internet, containing not only general
applications but also programs which only specialists need.  They are
provided in a way that users can easily install them on their
computer.  Users of special applications will not be troubled with
extra efforts like compiling the source themselves and follow a
complicated install procedure.

\subsubsection{Lost in the jungle of applications}

The largest collection of ready to run Free Software -- this sounds
great for marketing experts and makes a quite good statistics.  But
what would this mean for the unexperienced user who just started using
Debian and finished the first steps of the Debian installer?  Should
this poor user start reading the descriptions of 18,000 packages where
at best 50\% can be understand by a newcomer?

If this would really be the case the applications that are really
interesting for the user would be hidden perfectly and it would be
nearly the same as if they would be not packaged at all.  But there is
a light at the horizon.

\subsubsection{Custom Debian Distributions}

Custom Debian Distributions are completely integrated into Debian.  It
is a technique which enables a Debian user to focus on these parts of
Debian which are really interesting to solve a special task.  It is
implemented in so called ``meta packages'' which are simply spoken
packages which can only be installed if a set of other packages is
installed at the same time.  The packaging mechanisms inside Debian
care for everything if the user tries to install a meta package.

So in case a user is a teacher and wants to install a computer lab
with educational software for the students, focussing to the
\DebianEdu Custom Debian Distribution is the best idea.  In this
special case the situation has developed that far that there is even a
special CD to install Debian-Edu also called SkoleLinux and the whole
Debian system is not really needed, but in principle everything which
is on the SkoleLinux CD is included in Debian.

Another example is \DebianMed.  It was prepared for people working in
the field of health care.  There are a lot of applications that are
used for DNA or protein sequences inside Debian.  All of these will be
installed when one single package called \texttt{\em med-bio} is
installed.  The same procedure for medical imaging applications: The
installation of \texttt{\em med-imaging} causes the installation of all
packages that are relevant for users who have to deal with medical
imaging tasks.

This technique dispenses users that are working in special fields from
the task to do researches which applications do exist and which are
relevant for their own task.  The packages will not only be installed
at this computer but they also get a menu entry which is made visible
in a special user sub menu for easy access to the programs in
question.

\section*{Conclusion}

In the first it was explained why Free Software exists and that
``free'' in this sense does not really mean ``free of charge'' as free
beer.  In the second part also free knowledge (WikiPedia) was
explained and things people should know when dealing with Free
Software and their relation to proprietary software were mentioned.
Finally the third part explained which power is behind Free Software
if it is used sanely in a distribution which fits the needs of users.

\end{document}