[Cdd-commits] r1180 - projects/med/trunk/papers/09_osehc_porto
CDD Subversion Commit
noreply at alioth.debian.org
Fri Oct 31 08:42:19 UTC 2008
Author: tille
Date: Fri Oct 31 08:42:18 2008
New Revision: 1180
Added:
projects/med/trunk/papers/09_osehc_porto/paper-text_longversion.tex
Modified:
projects/med/trunk/papers/09_osehc_porto/paper-text.tex
Log:
The original paper was to long ... the maximum is 10 pages. :-(
Modified: projects/med/trunk/papers/09_osehc_porto/paper-text.tex
==============================================================================
--- projects/med/trunk/papers/09_osehc_porto/paper-text.tex (original)
+++ projects/med/trunk/papers/09_osehc_porto/paper-text.tex Fri Oct 31 08:42:18 2008
@@ -22,15 +22,6 @@
Edu, Debian Science and others.
\end{abstract}
-\subsubsection*{List of abbreviations}
-\begin{tabular}{ll} % to become llll later
-Blend & Debian Pure Blend \\
-BTS & Bug Tracking System \\
-DFSG & Debian Free Software Guidelines \\
-GUI & Graphical User Interface \\
-LSB & Linux Standard Base \\
-\end{tabular}
-
\section{Introduction}
@@ -61,27 +52,15 @@
on most computers and have a very large user base. Knowing this, many
gifted programmers feel obliged for this kind of Free Software - they
just need it for their own and they know to make a difference to the
-world. So one finds a fast, growing community around Free Software
-packages that have a widespread use. For specialised software in
-general and particularly for biological software, one needs to first
-explain what a particular software does. Often the developers get
-their satisfaction not from the software but from the excitement of
-extra insights in biological processes - the beauty of which is likely
-not to be accessible to regular software enthusiasts.
-
-Considering the fact that an important part of medical care is based
-on microbiology and genetic analysis this software forms one important
-column of the \DebianMed project. As a view shared by many in Open
-Source Bioinformatics, Ewan Birney of the EnsEMBL \cite{hubbard:2005}
-project stresses Open Source to "ensure scientific progress". He also
-laid out that the Open Source of programs is a comparatively trivial
-issue when compared with the openness of biological data. The latter
-is often far more expensive to produce. While giving the data away
-might possibly diminish one's competitive advantage, the sharing of a
-program with others and respective citations is well accepted as
-fostering one's career. This is one reason for the fact that today
-the field of microbiology is well covered by Free Software and the
-most important ones are now packaged for Debian.
+world.
+
+The fact that a piece of software is needed for the own work is often
+the basic motivation to write Free Software. Biological software is
+often developed by scientists themselves because they are the only
+persons who have the needed insight in the topic. Many authors of
+this software just realised the profit they might gain to share the
+code and thus the field of biology is really well covered with Free
+Software.
Another quite well covered part of \DebianMed is medical imaging.
Even if there are quite important packages like Bio Image Suite
@@ -159,23 +138,6 @@
of the technically skilled early adopters. Debian such helps to bring
developers and users of applications together.
-\begin{figure*}
-\centering
-\myinsertgraphic{earthkeyring}{80mm}{53mm}
-\caption{Debian web of trust}\label{figure:weboftrust}
-\end{figure*}
-
-All members of the Debian project are connected in a
-\printurl{people.debian.org/~edward/globe/earthkeyring}{web of trust},
-which is woven by signing GnuPG
-(\printurl{www.gnupg.org}{www.gnupg.org}) keys\ref{figure:weboftrust}.
-A central requirement to become a member of the Debian project is to
-have one's GPG key signed by an already accepted member of the Debian
-community. When Debian developers first meet in person, they sign
-each other's keys. Thus, the web of trust is woven.
-
-Debian does its best to have every member profit of somebody else's
-work as quickly as possible.
\subsubsection{Debian Policy}
@@ -184,26 +146,9 @@
(LSB)\cite{lsb} is attempting to develop and promote a set of
standards that will increase compatibility among Linux distributions,
hereby enabling software applications to run on any compliant system.
-The very essence of any distribution, (whether delivered as RPMs,
-DEBs, Source tarballs or ports) is the choice of {\itshape policy
+The very essence of any distribution is the choice of {\itshape policy
statements\/}.
-\halfsloppy{\itshape Policy statements\/} in Debian\cite{debianpolicy}
-specify configuration files to reside in
-\Path{/etc/\$package/\$package.conf}, logfiles go to
-\Path{/var/log/\$package/\$package.log} and the documentation files to
-be located in \Path{/usr/share/doc/\$package}. CGI-scripts are
-installed in \Path{/usr/lib/cgi-bin}.
-
-The policy statements are followed by the tool-chains and libraries
-used to build the software, and the lists of dependencies, which
-dictate the prerequisites and order in which the software has to be
-built and installed.
-
-Policies in Debian are developed within the community. Commonly with a
-single person or a small team drafting it, with further refinements
-being discussed in respective mailing lists.
-
While every single maintainer of a Debian package has to build the
package in compliance with the policy he has the ability and the right
to decide which software is worth packaging. Normally maintainers
@@ -269,16 +214,6 @@
\item Offering training and qualification.
\end{itemize}
-The best established Distributors of GNU/Linux systems are
-\printurl{www.mandrakelinux.com}{Mandrake},
-\printurl{www.redhat.com}{RedHat}, \printurl{www.suse.com}{SuSE} (now
-owned by \printurl{www.novell.com/linux}{Novell}) and
-\printurl{www.debian.org}{Debian}.
-\printurl{www.linspire.com}{Linspire},
-\printurl{www.xandros.com}{Xandros},
-\printurl{www.mepis.org}{MEPIS} and
-\printurl{www.ubuntulinux.org}{Ubuntu} are well known derivates of
-Debian.
\section{Results}
@@ -295,13 +230,6 @@
this site have been created by the EGTDC specifically for the
Bio-Linux project and are in deb format.
- The projects claim to be compatible with any installation of Debian
- or Debian variants. This might be a weak sentence because Debian
- variants are quite different. It is better to say exactly Debian
- "stable" or Debian "testing" and draw a line between native Debian
- and perhaps commercial derivatives which might differ in certain
- aspects.
-
The packages are not, however, policy compliant Debian packages
because they install files into \Path{/usr/local} hierarchy in
contrast to the Debian policy which does not allow files inside
@@ -312,11 +240,11 @@
Besides this technical fact Bio-Linux authors were not that strict
regarding licensing and copyright of packaged projects. Every
official Debian package has to comply to the Debian Free Software
- Guidelines DFSG and the copyright information has to be shipped with
- the file \Path{/usr/share/doc/<packagename>/copyright}. Moreover
- the source of a binary Debian package has to be provided next to the
- binary. All these very important requirements are not fulfilled in
- most cases in Bio-Linux.
+ Guidelines (DFSG) and the copyright information has to be shipped
+ with the file \Path{/usr/share/doc/<packagename>/copyright}.
+ Moreover the source of a binary Debian package has to be provided
+ next to the binary. All these very important requirements are not
+ fulfilled in most cases in Bio-Linux.
However, the authors did a great job in collecting a certain amount
of very useful software for biologists and the \DebianMed project
@@ -332,16 +260,6 @@
mainly for people who do not know to proceed unpacking a program,
compile and install it correctly.
- For these reasons, they want to give contribution, developing a
- Linux system that aims to be easy to use and still offering packages
- that will be part of the BioLinux-BR. Giving this option, we will
- not be compelling anybody to change its Linux distribution, since
- there are packages developed for the greater Linux distributions.
-
- "BR" here stands for "Brasil" which might lead to the assumption
- that some internationalisation effort was done, but according to the
- authors this was not their main concern but rather a side effect.
-
In fact, this project has assembled a huge amount of packages,
probably the most complete collection of Free Software in biology.
Packages for multiple distributions are provided, which includes
@@ -380,18 +298,8 @@
\item[The Quantian Scientific Computing Environment]
Quantian is a remastering of a well established effort
- (\printurl{www.knopper.net/knoppix/index-en.html}{Knoppix}). Recent
- versions of Quantian \cite{eddelbuettel:quantian} are based on
- \printurl{bofh.be/clusterknoppix/}{ClusterKnoppix
- }\cite{vandersmissen:2005} and add support for
- \printurl{www.openmosix.org}{OpenMosix} \cite{bar:openmosix},
- including remote booting of light clients in an openMosix terminal
- server context. Earlier releases are still available; see below for
- URLs for downloads as well as ordering information. More detailed
- information \printurl{dirk.eddelbuettel.com/quantian}{are
- available at the Quantian homepage}.
-
- The interesting part for biologists is that Quantian contains in
+ (\printurl{www.knopper.net/knoppix/index-en.html}{Knoppix}). The
+ interesting part for biologists is that Quantian contains in
addition all interesting packages of \DebianMed. The author Dirk
Eddelbuettel, who is a Debian developer himself, just used the
simply to install biological software feature we provide and thus
@@ -476,18 +384,17 @@
software to a large user base is a main advantage of \DebianMed.
The strength of Debian is the huge number of developers (more than
-1000) all over the world working in different
-fields\ref{figure:weboftrust}. Some of them are working in the field
-of biology or medicine and thus have a natural interest in developing
-a rock solid system they can relay on for their own work (not only
-commercial interest to sell service per accident). So sometimes the
-chances to realise specific support for small user groups are better
-inside a community driven distribution than in a commercial
-distribution: You just need some developers who have a specific
-interest and they will realise and publish an environment for their
-needs and will share it with other users. A company that has to gain
-a certain market share is not flexible enough in this regard to cover
-very specific interests.
+1000) all over the world working in different fields. Some of them
+are working in the field of biology or medicine and thus have a
+natural interest in developing a rock solid system they can relay on
+for their own work (not only commercial interest to sell service per
+accident). So sometimes the chances to realise specific support for
+small user groups are better inside a community driven distribution
+than in a commercial distribution: You just need some developers who
+have a specific interest and they will realise and publish an
+environment for their needs and will share it with other users. A
+company that has to gain a certain market share is not flexible enough
+in this regard to cover very specific interests.
The underlying principle that those things will be done if there is
somebody who just does the work is called {\em Do-O-Cracy} -- which
@@ -528,11 +435,10 @@
user types in: \\
\hspace*{10mm}\texttt{apt-get install med-bio} \\
all applications inside Debian which are related to the field of
-molecular biology and medical genetics will be installed.The
+molecular biology and medical genetics will be installed. The
\package{med-bio-dev} package just installs programming libraries and
tools which are interesting for users who want to develop biological
-applications (for instance the NCBI
-library\cite{Wheeler2005,Jenuth2000,ncbi:2004}).
+applications.
\subsubsection{Continuous growth}
@@ -568,15 +474,14 @@
\caption{Activity of most active authors on the \DebianMed mailing list}\label{figure:authorstats}
\end{figure*}
-\enlargethispage{10mm}
-This growth of the output of a project is an important part but
-we also try to measure the commitment of the people involved in the
+This growth of the output of a project is an important part but we
+also try to measure the commitment of the people involved in the
project. It has to be ensured that fresh blood is flooding into the
project to make sure it can cope with the normal loss of supporters
which always happens in Free Software projects (people find new jobs
with different orientation or less spare time for private reasons
etc.) A raw measure for the activity of members might be their mails
-to the project mailing list. Figure \ref{figure:authorstats}) shows
+to the project mailing list. Figure \ref{figure:authorstats}) shows
the number of mails of the ten most active posters of the \DebianMed
mailing list. This graph shows perfectly that the number of active
supporters is growing solidly. In the last two years three people can
@@ -640,57 +545,6 @@
of ready-to-install Free Software on the Internet.
-\subsubsection{Licensing issues}
-
-Several existing programs that might be useful for specialists are not
-free in the sense of the
-\printurl{www.debian.org/social\_contract\#guidelines}{Debian Free
- Software Guidelines} \cite{dfsg}. Programs that are incompatible
-with the DFSG cannot be included in the Debian main distribution. One
-famous example of this group is
-\printurl{evolution.genetics.washington.edu/phylip.html}{PhyLip} and
-the same hold for ClustalW. Both programs' licenses contain a clause
-like
-\begin{quote}
- Permission is granted to copy and use this program provided no fee
- is charged for it and provided that this copyright notice is not
- removed.
-\end{quote}
-As a consequence no reseller of Debian would be allowed to sell Debian
-because one CD contains PhyLip. Hence, this program may not be
-redistributed per default with the main and essential tools of
-Debian. Problems are, starting with those for the user:
-\begin{itemize}
-\item Need to obtain PhyLip from a different source
-\item If the user is not using a "common" architecture like {\ttfamily
- i386} he might run into problems in compiling the latest version
- as only the packages of the main distribution are submitted to the
- build daemons to save resources.
-%\item No extra security checks from his trusted operating system vendor.
-\item The user might not even notice that something like PhyLip exists
- at all. \DebianMed cares for pointing users to relevant software
- and thus the user will be pointed to each single program package
- {\itshape inside\/} Debian which of interest for biological
- research.
-\end{itemize}
-
-%While Debian Med supports packages in the non-free section of
-%Debian this might be dropped at any time and has certain constraints.
-From the developers point of view we face also drawbacks:
-\begin{itemize}
-\item Possibly smaller user base (see last point above)
-\item Fewer bug reports and thus lower chance to increase the quality
- of their software.
-\item \halfsloppy Porting problems to different architectures might
- not be revealed early.
-\end{itemize}
-
-While every author is perfectly free to choose this kind of license,
-Free Software experts agree that this kind of restrictions is possibly
-a drawback for those programs because they do not fully use the
-spinning power of Free Software development.
-
-
\subsubsection{Importance of community support}
That strong support within the community of users is essential for the
@@ -718,7 +572,7 @@
that was done -- it contains also a nicely formated list which
projects would be interesting to reach the final goal to cover each
task in medical care with Free Software. This list inside each
- section might give some idea where help might be needede.
+ section might give some idea where help might be needed.
\item \halfsloppy Verify whether one needs special configuration for
your project. If yes, verify which possibilities are given in the
Debian Pure Blends effort. It is more than collecting software but
Added: projects/med/trunk/papers/09_osehc_porto/paper-text_longversion.tex
==============================================================================
--- (empty file)
+++ projects/med/trunk/papers/09_osehc_porto/paper-text_longversion.tex Fri Oct 31 08:42:18 2008
@@ -0,0 +1,773 @@
+\newcommand{\DebianMed}{Debian\ Med\xspace}
+
+\begin{abstract}
+ The Debian Med project started in 2002 with the objective to bring
+ free medical software into the focus of users. The first step was
+ the investigation of available Free Software in this area and to
+ accomplish the conditions for a simple and solid installation of
+ this software in Debian GNU/Linux. It has shown that the Debian Med
+ project has a positive effect for the cooperating upstream projects.
+
+ It has shown that fields like molecular biology and medical imaging
+ were covered quite good with Free Software solutions while there is
+ not really much to manage a medical practice. The Debian Med project
+ tries to be the missing link between developers and users to support
+ Free Software solutions for all areas in medicine.
+
+ Debian Med is intended to be useful for service providers who want
+ to distribute their free or proprietary solutions in medical care.
+
+ The whole project is embeded into the Debian Pure Blends framework
+ (formerly known as Custom Debian Distributions) as well as Debian
+ Edu, Debian Science and others.
+\end{abstract}
+
+\subsubsection*{List of abbreviations}
+\begin{tabular}{ll} % to become llll later
+Blend & Debian Pure Blend \\
+BTS & Bug Tracking System \\
+DFSG & Debian Free Software Guidelines \\
+GUI & Graphical User Interface \\
+LSB & Linux Standard Base \\
+\end{tabular}
+
+
+\section{Introduction}
+
+\subsection{Motivation and purpose of \DebianMed}
+
+The advent of \DebianMed is characterised by the rise of several new
+Free Software projects that were useful for certain tasks in medical
+care. Several of these projects -- even if looking promising in the
+beginning -- do not really exist any more because they did not
+respected the principle of Free Software: care for a solid user base
+and recruit qualified developers from it. This task is one of the
+hardest in the field of health care: there is a much smaller number of
+users for such specific software compared to a web browser or an
+office suite.
+
+A big problem was that it was quite hard to obtain solid information
+about all these projects. Some engaged users tried to assemble lists
+of these projects and published these in the web. While those lists
+are an interesting start they are finally not really helpful for
+users: Users just need ready to run programs with no effort to obtain
+and install the software. The idea of \DebianMed was now to provide
+not only a list but exactly what users need.
+
+
+\subsection{Status of Free Software in health care}
+
+Common programs like a web server, or a mail user agent are installed
+on most computers and have a very large user base. Knowing this, many
+gifted programmers feel obliged for this kind of Free Software - they
+just need it for their own and they know to make a difference to the
+world. So one finds a fast, growing community around Free Software
+packages that have a widespread use. For specialised software in
+general and particularly for biological software, one needs to first
+explain what a particular software does. Often the developers get
+their satisfaction not from the software but from the excitement of
+extra insights in biological processes - the beauty of which is likely
+not to be accessible to regular software enthusiasts.
+
+Considering the fact that an important part of medical care is based
+on microbiology and genetic analysis this software forms one important
+column of the \DebianMed project. As a view shared by many in Open
+Source Bioinformatics, Ewan Birney of the EnsEMBL \cite{hubbard:2005}
+project stresses Open Source to "ensure scientific progress". He also
+laid out that the Open Source of programs is a comparatively trivial
+issue when compared with the openness of biological data. The latter
+is often far more expensive to produce. While giving the data away
+might possibly diminish one's competitive advantage, the sharing of a
+program with others and respective citations is well accepted as
+fostering one's career. This is one reason for the fact that today
+the field of microbiology is well covered by Free Software and the
+most important ones are now packaged for Debian.
+
+Another quite well covered part of \DebianMed is medical imaging.
+Even if there are quite important packages like Bio Image Suite
+\cite{bioimagesuite:2008} not yet packaged for Debian first steps are
+done and there is a certain coverage of DICOM viewers and other
+software which is very useful in medical imaging.
+
+It has shown that the field of patient management and medical health
+record applications which is considered by most people as the basic
+health care software has only one single representative which is
+GNUmed \cite{gnumed:2008}. It has turned out that there are a lot of
+similar Free Software projects that try to solve more or less the same
+problem but diverge in the techniques used (programming language,
+database server, etc.), user interface (GUI or web aplication) and the
+basic ideas about workflow and philosophy.
+
+The even larger task to manage a hospital is also tackled by several
+projects and the most famous one OpenVista\cite{openvista:2008} as a
+real enterprise grade health care information system is a really
+complex project and needs a stronger team than the current \DebianMed
+team with a larger technical background in the specific techniques
+used in OpenVista. To overcome this problem the strategy of
+\DebianMed is to try to involve the upstream authors into the
+packaging: They are really the experts of the software in question and
+we try to teach and help them in packaging.
+
+
+\section{Methods}
+
+\subsection{Unique technology supporting Debian's principles in packaging}
+
+\subsubsection{Build daemons}
+
+Sites external to the Debian main distribution may offer packages only
+for a subset of architectures. The Debian main distribution, however,
+automatically compiles software for all 11 architectures that are
+supported by the Debian effort. To get a package into Debian, be it
+novel or an update of an existing package, the maintainer of a package
+submits the source code of the program together with his changes on
+the code to create the package. The build daemons (or autobuilders)
+compile the packages for each of the supported systems and make the
+resulting package publicly available for download. Logs of the build
+platforms are available online for everybody's inspection.
+
+
+\subsubsection{Bug tracking system}
+
+Users should give immediate feedback about problems arising in using a
+package. They always have the choice of reporting these to the
+upstream developer, usually per email. A particular strength of
+\printurl{www.sourceforge.net}{SourceForget.net} is to bring users of
+a particular software together. Earlier than this effort was the
+\printurl{www.debian.org/Bugs}{Debian Bug Tracking System} (BTS). The
+maintainer of a software can decide if the bug should be forwarded to
+the upstream developers of the package or if it is fixed by
+himself. All problems are made public and hence the whole community
+may contribute to solving a particular issue.
+
+
+\subsection{Sharing the work}
+
+The Debian Project is an association of individuals who share the
+intention to create the best possible free operating system. This
+operating system that is created is called Debian GNU/Linux, or simply
+Debian for short. Everybody in the internet may initiate a site and
+offer packages for the installation in Debian. A local administrator
+has to decide, if this public source may be trusted.
+
+For Free Software development to work it requires a critical mass of
+supporters. Development without feedback prior to the submission of
+the final product is disadvantageous. The development of programs is
+not the main concern of a regular Linux Distribution. However, with
+the focus on Free Software and smooth local compilation, Debian
+considerably eases the contribution of comments and concise feedback
+of the technically skilled early adopters. Debian such helps to bring
+developers and users of applications together.
+
+\begin{figure*}
+\centering
+\myinsertgraphic{earthkeyring}{80mm}{53mm}
+\caption{Debian web of trust}\label{figure:weboftrust}
+\end{figure*}
+
+All members of the Debian project are connected in a
+\printurl{people.debian.org/~edward/globe/earthkeyring}{web of trust},
+which is woven by signing GnuPG
+(\printurl{www.gnupg.org}{www.gnupg.org}) keys\ref{figure:weboftrust}.
+A central requirement to become a member of the Debian project is to
+have one's GPG key signed by an already accepted member of the Debian
+community. When Debian developers first meet in person, they sign
+each other's keys. Thus, the web of trust is woven.
+
+Debian does its best to have every member profit of somebody else's
+work as quickly as possible.
+
+\subsubsection{Debian Policy}
+
+All GNU/Linux distributions have a certain amount of common ground,
+and the \printurl{www.linuxbase.org/}{Linux Standard Base}
+(LSB)\cite{lsb} is attempting to develop and promote a set of
+standards that will increase compatibility among Linux distributions,
+hereby enabling software applications to run on any compliant system.
+The very essence of any distribution, (whether delivered as RPMs,
+DEBs, Source tarballs or ports) is the choice of {\itshape policy
+ statements\/}.
+
+\halfsloppy{\itshape Policy statements\/} in Debian\cite{debianpolicy}
+specify configuration files to reside in
+\Path{/etc/\$package/\$package.conf}, logfiles go to
+\Path{/var/log/\$package/\$package.log} and the documentation files to
+be located in \Path{/usr/share/doc/\$package}. CGI-scripts are
+installed in \Path{/usr/lib/cgi-bin}.
+
+The policy statements are followed by the tool-chains and libraries
+used to build the software, and the lists of dependencies, which
+dictate the prerequisites and order in which the software has to be
+built and installed.
+
+Policies in Debian are developed within the community. Commonly with a
+single person or a small team drafting it, with further refinements
+being discussed in respective mailing lists.
+
+While every single maintainer of a Debian package has to build the
+package in compliance with the policy he has the ability and the right
+to decide which software is worth packaging. Normally maintainers
+choose the software which is used in their own work and they are free
+to move the development of Debian in a certain direction (as long as
+they follow the rules of the policy). This is referred to as {\em
+ Do-o-cracy} in Debian which means: The doer decides what is done.
+
+\subsection{Selection of packages}
+
+Debian contains nearly 20000 binary packages, and this number is
+constantly increasing. There is no single user who needs all these
+packages. The regular user is interested in a subset of these
+packages. To specify packages of one's particular interest, several
+options are provided by Debian:
+\begin{description}
+\item[tasksel] Provision of a reasonable selection of rather general
+ tasks that can be accomplished using a set of packages installed on
+ a Debian GNU/Linux system. However, these are not yet covering
+ scientific applications. The Blend toolkit which is currently
+ developed will also support \command{tasksel} to enable selecting
+ for instance Debian Med right after a fresh installation of a
+ general Debian system.
+
+\item[command line package management] \command{apt} provide means to
+ search for packages of particular interest by its name or words in
+ the package's description. Every package also indicates, as set by
+ its maintainer, references to other packages of potential interest.
+
+\item[GUI] There are several graphical user interfaces to manipulate
+ the installation of packages on a Debian installation. The most
+ popular is currently \command{synaptic} which eneables users to seek
+ for certain packages and displays detailed information about each
+ package.
+\end{description}
+
+Debian officially maintains 11 different architectures with many more
+not officially supported ports to other operating systems, which
+includes some that run another flavour of UNIX. Its technology for
+package management has been adopted for other operating systems,
+i.e. Fink on MacOSX
+(\printurl{fink.sourceforge.net}{fink.sourceforge.net}).
+
+A {\itshape distribution\/} is a collection of software packages
+around the GNU Linux operating system that satisfies the needs of the
+target user group. There are general distributions, which try to
+support all users, and there are several specialised distributions,
+which each target a special group of users.
+
+{\itshape Distributors\/} are those companies that are building these
+collections of software around the GNU Linux operating system. Since
+the software is Free, the user who buys a distribution pays for the
+service that the distributor is providing. These services might be:
+\begin{itemize}
+ \item Preparing a useful collection of software around GNU Linux.
+ \item Caring for smooth installation that the target user is able to
+ manage.
+ \item Providing software updates and security fixes.
+ \item Writing documentation and translations to enable the user to
+ use the distribution with maximum effect.
+ \item Selling Boxes with ready to install CDs and printed
+ documentation.
+ \item Offering training and qualification.
+\end{itemize}
+
+The best established Distributors of GNU/Linux systems are
+\printurl{www.mandrakelinux.com}{Mandrake},
+\printurl{www.redhat.com}{RedHat}, \printurl{www.suse.com}{SuSE} (now
+owned by \printurl{www.novell.com/linux}{Novell}) and
+\printurl{www.debian.org}{Debian}.
+\printurl{www.linspire.com}{Linspire},
+\printurl{www.xandros.com}{Xandros},
+\printurl{www.mepis.org}{MEPIS} and
+\printurl{www.ubuntulinux.org}{Ubuntu} are well known derivates of
+Debian.
+
+\section{Results}
+
+\subsection{Comparable Debian-associated repositories}
+
+\begin{description}
+
+\item[Bio-Linux Bioinformatics package repository]
+
+ \halfsloppy The Bio-Linux Bioinformatics package repository contains
+ the \printurl{envgen.nox.ac.uk/pkg\_repository.html}{Bio-Linux 4
+ bioinformatics software} and can be installed from a centralised
+ repository located on the EGTDC server. The packages available from
+ this site have been created by the EGTDC specifically for the
+ Bio-Linux project and are in deb format.
+
+ The projects claim to be compatible with any installation of Debian
+ or Debian variants. This might be a weak sentence because Debian
+ variants are quite different. It is better to say exactly Debian
+ "stable" or Debian "testing" and draw a line between native Debian
+ and perhaps commercial derivatives which might differ in certain
+ aspects.
+
+ The packages are not, however, policy compliant Debian packages
+ because they install files into \Path{/usr/local} hierarchy in
+ contrast to the Debian policy which does not allow files inside
+ packages at this location because \Path{/usr/local} is reserved for
+ locally installed files that do not fall under responsibility of the
+ Debian package manager.
+
+ Besides this technical fact Bio-Linux authors were not that strict
+ regarding licensing and copyright of packaged projects. Every
+ official Debian package has to comply to the Debian Free Software
+ Guidelines DFSG and the copyright information has to be shipped with
+ the file \Path{/usr/share/doc/<packagename>/copyright}. Moreover
+ the source of a binary Debian package has to be provided next to the
+ binary. All these very important requirements are not fulfilled in
+ most cases in Bio-Linux.
+
+ However, the authors did a great job in collecting a certain amount
+ of very useful software for biologists and the \DebianMed project
+ is seeking for possibilities for cooperation.
+
+\item[BioLinux-BR Project]
+
+ \halfsloppy A similar project is the
+ \printurl{biolinux.df.ibilce.unesp.br/index.en.php}{BioLinux-BR
+ Project} which is a project directed to the scientific community.
+ Their goal is to create a Linux distribution for people with little
+ familiarity with the installation of the operational system and
+ mainly for people who do not know to proceed unpacking a program,
+ compile and install it correctly.
+
+ For these reasons, they want to give contribution, developing a
+ Linux system that aims to be easy to use and still offering packages
+ that will be part of the BioLinux-BR. Giving this option, we will
+ not be compelling anybody to change its Linux distribution, since
+ there are packages developed for the greater Linux distributions.
+
+ "BR" here stands for "Brasil" which might lead to the assumption
+ that some internationalisation effort was done, but according to the
+ authors this was not their main concern but rather a side effect.
+
+ In fact, this project has assembled a huge amount of packages,
+ probably the most complete collection of Free Software in biology.
+ Packages for multiple distributions are provided, which includes
+ Debian, and a live CD.
+
+\end{description}
+
+\subsection{Other repositories of biology related software}
+
+Looking beyond Debian and related distributions which share more or
+less the same technique we find similar efforts to deliver sets of
+ready to install software
+
+\begin{description}
+
+\item[FreeBSD Ports: Biology] The Free Software world does not only
+ know Linux as free operating system. There are others out there
+ like several BSD derivates, OpenSolaris, Hurd and others. The
+ FreeBSD project has a really nice
+ \printurl{http://www.freebsd.org/ports/biology.html}{collection of
+ biological software}.
+\end{description}
+
+\subsection{Bioinformatics Live CDs}
+
+The concept of a live CD allows to create a CD or DVD that boots a
+computer, starts a defined set of application without a user's
+intervention and has all tools in place that suits a particular
+community. Such provide fully featured Linux workstations without
+additional installations of access to local disk space, alternatively
+booting via the network is supported by Debian, which particularly
+appeals to Blades or large clusters. The most successful such LiveCD
+is the Debian-derived Knoppix\cite{knopper:2005}.
+
+\begin{description}
+\item[The Quantian Scientific Computing Environment]
+
+ Quantian is a remastering of a well established effort
+ (\printurl{www.knopper.net/knoppix/index-en.html}{Knoppix}). Recent
+ versions of Quantian \cite{eddelbuettel:quantian} are based on
+ \printurl{bofh.be/clusterknoppix/}{ClusterKnoppix
+ }\cite{vandersmissen:2005} and add support for
+ \printurl{www.openmosix.org}{OpenMosix} \cite{bar:openmosix},
+ including remote booting of light clients in an openMosix terminal
+ server context. Earlier releases are still available; see below for
+ URLs for downloads as well as ordering information. More detailed
+ information \printurl{dirk.eddelbuettel.com/quantian}{are
+ available at the Quantian homepage}.
+
+ The interesting part for biologists is that Quantian contains in
+ addition all interesting packages of \DebianMed. The author Dirk
+ Eddelbuettel, who is a Debian developer himself, just used the
+ simply to install biological software feature we provide and thus
+ made a great profit from \DebianMed.
+
+\item[Vigyaan - the biochemical software workbench]
+
+ \halfsloppy\printurl{www.vigyaancd.org}{Vigyaan} is an electronic
+ workbench for bioinformatics, computational biology and
+ computational chemistry. It has been designed to meet the needs of
+ both beginners and experts. VigyaanCD is a live Linux CD containing
+ all the required software to boot the computer with ready to use
+ modelling software. VigyaanCD v0.1 is based on Knoppix v3.3.
+
+ Vigyann contains some programs which are not yet contained in
+ Debian. It might be mutually beneficial to include these provided
+ that the license fits the DFSG.
+
+\item[BioKnoppix]
+
+ \printurl{bioknoppix.hpcf.upr.edu}{BioKnoppix} is a customised
+ distribution of Knoppix Linux Live CD. It is a very similar project
+ to the previous which specialises Knoppix for computational biology
+ and chemistry.
+
+\item[VLinux Bioinformatics Workbench]
+
+ Also \printurl{bioinformatics.org/vlinux}{VLinux} is at the time of
+ writing a Live CD based on the same outdated Knoppix version 3.3 as
+ Vigyann and includes a slightly changed software selection and
+ surely a different background layout.
+
+\end{description}
+
+These are too many different initiatives that could all well do much
+more in order to share the burden of maintenance and updates. With
+Debian they have he right basic infrastructure. The time will show,
+whose packages will gain most momentum.
+
+
+\subsection{Comparison with \DebianMed}
+
+\subsubsection{Other fields than only biology}
+
+Looking at all the projects above it becomes evident that they are all
+dealing only with biological software. Above it was stated that one
+strong column of \DebianMed is this specific field and it is for a
+reason: The amount of free biological software is large and most of
+these projects are relatively easy to turn into packages - so the
+amount of work per package is much smaller compared to for instance
+medical record applications with preparation of databases, dedicated
+user management, etc.
+
+So there is one major difference between the projects mentioned above
+and \DebianMed: While the biological part is really interesting for
+medical care \DebianMed tries to cover all other fields of medical
+care as well. This goal is not yet reached but continuous work is
+done into this direction and some important steps are done.
+
+\subsubsection{Debian Pure Blend}
+
+An even more important difference than the more general approach
+compared to the other repositories is the fact that \DebianMed is not
+only about just packaging software. The \DebianMed project is one of
+the earliest so called {\em Debian Pure Blends} (formerly known as
+Custom Debian Distributions) and just wants to do more for the comfort
+of their users than adding binary packages to the Debian package pool.
+The main goal is to turn Debian into the distribution of choice for
+people working in the field of medicine and to make Debian an
+operating system that is particularly well fit for the requirements
+for medical practice and research. The goal of \DebianMed is a
+complete system for all tasks in medical care which is build
+completely on free software.
+
+
+\subsubsection{Flexibility in supporting small user groups}
+
+On the organisational side the project tries to attract people working
+in the field of Free Software in medicine to share the effort of
+building a common platform which reduces the amount of work for
+developers and users. Sharing the effort in publishing free medical
+software to a large user base is a main advantage of \DebianMed.
+
+The strength of Debian is the huge number of developers (more than
+1000) all over the world working in different
+fields\ref{figure:weboftrust}. Some of them are working in the field
+of biology or medicine and thus have a natural interest in developing
+a rock solid system they can relay on for their own work (not only
+commercial interest to sell service per accident). So sometimes the
+chances to realise specific support for small user groups are better
+inside a community driven distribution than in a commercial
+distribution: You just need some developers who have a specific
+interest and they will realise and publish an environment for their
+needs and will share it with other users. A company that has to gain
+a certain market share is not flexible enough in this regard to cover
+very specific interests.
+
+The underlying principle that those things will be done if there is
+somebody who just does the work is called {\em Do-O-Cracy} -- which
+just means the doer decides what gets done.
+
+That is the reason why Debian is often the platform of choice for
+researcher in the field of biology: Some biologists are Debian
+maintainers and so they added support for biological packages. The
+more the Debian user in the field of biology report back about
+problems or wishes the more Debian maintainers are able to enhance
+their system for their own and their users profit.
+
+\subsubsection{Metapackages}
+
+On the technical side \DebianMed contains a set of metapackages that
+declare dependencies on other Debian packages, and that way the
+complete system is prepared for solving particular tasks. So the user
+has not to deal with the large number of package descriptions of 20000
+packages inside the Debian distribution -- it is just enough to seek
+for metapackages starting with prefix \package{med-} and install the
+metapackage of choice. The package management system will care for
+the installation of all packages that are in the list of dependencies
+of this metapackage - so the user can be sure that all packages he
+might need for the job will be installed on his system. Once one of
+the metapackages is installed a special user menu will be created to
+enhance usability for the user working in the field of medicine.
+
+Currently inside \DebianMed applications are provided in certain
+categories: medical practice and patient management, medical research,
+hospital information systems, medical imaging, documentation,
+molecular biology and medical genetics and others.
+
+There are two so called metapackages which are named \package{med-bio}
+and \package{med-bio-dev}. The sense of a meta package is that you
+have to install only one single package using a package management
+software inside Debian to get all interesting packages which are
+necessary for a single task. For instance if a
+user types in: \\
+\hspace*{10mm}\texttt{apt-get install med-bio} \\
+all applications inside Debian which are related to the field of
+molecular biology and medical genetics will be installed.The
+\package{med-bio-dev} package just installs programming libraries and
+tools which are interesting for users who want to develop biological
+applications (for instance the NCBI
+library\cite{Wheeler2005,Jenuth2000,ncbi:2004}).
+
+
+\subsubsection{Continuous growth}
+
+Several Free Software projects which try to deal with small user group
+software started with a lot of enthusiasm but at some point in time
+developers had other interests or just were unable to maintain the
+project because of lack of man power. The strategy of \DebianMed is
+to stay strictly inside Debian -- so even if manpower is a problem the
+whole infrastructure around will stay solid and does not drain extra
+resources. So nobody of the \DebianMed team has to care about writing
+installers, running an online repository and mirrors work on a bug
+tracking system etc. All this infrastructure is just there.
+
+\begin{figure*}
+\centering
+\myinsertgraphic{dmstats.pdf}{120mm}{80mm}
+\caption{Number of dependencies of selected metapackages}\label{figure:dmstats}
+\end{figure*}
+
+The success of this strategy can be proven by a continuous growth for
+instance if the number of packages inside Debian which is interesting
+for health care. Taking the number of dependencies of some
+metapackages into account (see figure \ref{figure:dmstats}) at the
+beginning of the project in 2002 a quite low number of packages useful
+for medical care was available. A nearly linear growth with a
+gradient that perfectly reflects the availability of programs in this
+field can be observed.
+
+\begin{figure*}
+\centering
+\myinsertgraphic{authorstat.pdf}{120mm}{80mm}
+\caption{Activity of most active authors on the \DebianMed mailing list}\label{figure:authorstats}
+\end{figure*}
+
+\enlargethispage{10mm}
+This growth of the output of a project is an important part but
+we also try to measure the commitment of the people involved in the
+project. It has to be ensured that fresh blood is flooding into the
+project to make sure it can cope with the normal loss of supporters
+which always happens in Free Software projects (people find new jobs
+with different orientation or less spare time for private reasons
+etc.) A raw measure for the activity of members might be their mails
+to the project mailing list. Figure \ref{figure:authorstats}) shows
+the number of mails of the ten most active posters of the \DebianMed
+mailing list. This graph shows perfectly that the number of active
+supporters is growing solidly. In the last two years three people can
+be considered as very active and there is no dramatic loss of people
+in the project. The general activity on the mailing list is
+constantly growing.
+
+
+\subsubsection{Role inside Debian}
+
+The Debian Pure Blends framework was mainly inspired by the \DebianMed
+needs. Regarding implementation the Debian Edu project -- another
+Blend -- is much more advanced. There are several reasons for this
+fact. The main reasons are the availability of software for education
+and the fact that a Norwegian company payed developers to work full
+day on this project. \DebianMed tried to gain profit from common
+technologies and generalised some tools of Debian Edu for all Blends.
+Moreover inside \DebianMed some new Blend tools were developed which
+enhance the easy build of metapackages. The latest contribution to
+the Blend framework is a set of automatically generated web pages
+which can be used by any Blend. For instance there is an overview
+about all the so called tasks (fields like biology, medical imaging,
+practice management) which provides a nice list of all the packages
+including some metainformation and the description of the packages in
+this task. So the output of the project can be easily viewed at
+\printurl{http://debian-med.alioth.debian.org/tasks/}{the tasks page
+ of \DebianMed}.
+
+By providing this kind of tools for other Blends as well \DebianMed
+has set some cornerstone in the technical cooperation between user
+oriented projects inside Debian. This in turn awakes the interest of
+other Debian developers who might provide other tools for \DebianMed.
+
+\section{Discussion}
+
+The \DebianMed project serves as a common platform for all Free
+Software that may be utilised in medical care. Tools developed in
+computational biology is just a part of it because it is an important
+brick in medical science. With \DebianMed's ambition to become the
+platform of choice for medical work, conform with the principles of
+the \printurl{www.debian.org/doc/debian-policy}{Debian
+ Policy}\cite{debianpolicy}, by the means of the distribution of
+development within the Debian Society, a well established reference
+platform for bioinformatics research and its medical applications has
+evolved and will continue to improve. The organisation is open, both
+to new members and to external sites offering packages for
+installations.
+
+
+\subsection{Differences from other distributions}
+
+The Debian GNU/Linux distribution differs from others in several
+ways. Firstly, Debian is a non-commercial organisation of volunteers,
+that does not sell anything. The second and most appealing difference
+is the peer review and continuous pressure among the members to
+provide a high quality of packages. The Debian society has a
+constitution, elects its leader, and transparently describes policies
+for the creation of packages utilising specific technologies.
+
+With these principles, Debian achieved the largest collection
+of ready-to-install Free Software on the Internet.
+
+
+\subsubsection{Licensing issues}
+
+Several existing programs that might be useful for specialists are not
+free in the sense of the
+\printurl{www.debian.org/social\_contract\#guidelines}{Debian Free
+ Software Guidelines} \cite{dfsg}. Programs that are incompatible
+with the DFSG cannot be included in the Debian main distribution. One
+famous example of this group is
+\printurl{evolution.genetics.washington.edu/phylip.html}{PhyLip} and
+the same hold for ClustalW. Both programs' licenses contain a clause
+like
+\begin{quote}
+ Permission is granted to copy and use this program provided no fee
+ is charged for it and provided that this copyright notice is not
+ removed.
+\end{quote}
+As a consequence no reseller of Debian would be allowed to sell Debian
+because one CD contains PhyLip. Hence, this program may not be
+redistributed per default with the main and essential tools of
+Debian. Problems are, starting with those for the user:
+\begin{itemize}
+\item Need to obtain PhyLip from a different source
+\item If the user is not using a "common" architecture like {\ttfamily
+ i386} he might run into problems in compiling the latest version
+ as only the packages of the main distribution are submitted to the
+ build daemons to save resources.
+%\item No extra security checks from his trusted operating system vendor.
+\item The user might not even notice that something like PhyLip exists
+ at all. \DebianMed cares for pointing users to relevant software
+ and thus the user will be pointed to each single program package
+ {\itshape inside\/} Debian which of interest for biological
+ research.
+\end{itemize}
+
+%While Debian Med supports packages in the non-free section of
+%Debian this might be dropped at any time and has certain constraints.
+From the developers point of view we face also drawbacks:
+\begin{itemize}
+\item Possibly smaller user base (see last point above)
+\item Fewer bug reports and thus lower chance to increase the quality
+ of their software.
+\item \halfsloppy Porting problems to different architectures might
+ not be revealed early.
+\end{itemize}
+
+While every author is perfectly free to choose this kind of license,
+Free Software experts agree that this kind of restrictions is possibly
+a drawback for those programs because they do not fully use the
+spinning power of Free Software development.
+
+
+\subsubsection{Importance of community support}
+
+That strong support within the community of users is essential for the
+development of software, for quality assurance, feedback on features,
+and not at least for the motivation of staff, all commercial
+distributors are well aware of. E.g., RedHat has initiated Fedora as
+a free supplement to their commercial distribution. It is this reason
+why \DebianMed is part of Debian and why groups external to the Debian
+society, like BioLinux, are also keen on close collaborations with the
+community.
+
+
+\subsubsection{Road map to come closer to \DebianMed}
+
+\begin{enumerate}
+\item Join the \printurl{lists.debian.org/debian-med}{\DebianMed
+ mailing list}.
+\item Check what projects are missing and ask Debian maintainers for
+ official inclusion. There is a sponsoring program by which even non
+ Debian developers can provide packages which are checked and
+ uploaded by official maintainers. There is no point in keeping good
+ quality softwares outside of Debian.
+\item \printurl{http://debian-med.alioth.debian.org/tasks/}{The tasks
+ page of \DebianMed} mentioned above does not only contain the work
+ that was done -- it contains also a nicely formated list which
+ projects would be interesting to reach the final goal to cover each
+ task in medical care with Free Software. This list inside each
+ section might give some idea where help might be needede.
+\item \halfsloppy Verify whether one needs special configuration for
+ your project. If yes, verify which possibilities are given in the
+ Debian Pure Blends effort. It is more than collecting software but
+ bringing the software to your target users while taking the burden
+ from any configuration issues from his back.
+\item The only reason to keep things outside of Debian are licenses
+ which are not compatible with DFSG. All other parts of your
+ projects can be included and your time for everyday package building
+ tasks can be saved and the workload shared with other people
+ following the same road.
+\end{enumerate}
+
+
+\section{Conclusions}
+
+We have shown that there is a considerable heterogeneous shape of Free
+Software for medical care even if some fields like micro biology are
+better covered than others. The continuous updates of data and the
+addition of novel important tools for a general medical environment
+cannot be performed by a single maintainer. The adherence to a policy
+and the sharing of maintenance are basic technologies to allow
+inter-institutional software projects of different kind in health
+care.
+
+Debian and its special dedication to medical software in \DebianMed,
+but also the technical infrastructure behind this community project
+renders a comfortable solution. The volunteers behind \DebianMed
+strive to support everybody's specific projects as best as they
+possibly can. It is the particular challenge of users of Free
+Software, to determine together with the community the available
+packages that already serve their needs or may be adapted
+respectively.
+
+For Debian GNU/Linux to become the race-horse for Free Software in
+health care, further important software which is listed at
+\printurl{http://en.wikipedia.org/wiki/List_of_open_source_healthcare_software}{Wikipedia
+ list of open source healthcare software}\cite{wikipediaoslist:2008}
+like for instance OpenVista\cite{openvista:2008} and other enterprise
+grade health care information systems has to be packaged for Debian.
+
+Whatever choice for an infrastructure is made, with Debian it is
+available to all collaborating sites almost instantly.
+
+\subsection*{Acknowledgements}
+
+The author thanks the Debian Developers for all their work and
+particularly those who have contributed to \DebianMed and the Debian
+Pure Blends effort which builds a common framework around similar
+projects like Debian Edu for education, Debian Science for scientific
+work and others. Debian is supported by donations through
+\printurl{www.spi-inc.org}{Software in the Public Interest}, a
+non-profit umbrella organisation for free software projects.
More information about the Cdd-commits
mailing list