Talk at LinuxTag in Karlsruhe
June 7th, 2002
Talk about the two internal Debian Projects Debian Med and Debian-Junior at LinuxTag in Karlsruhe. The complete materials are available as MagicPoint source and preformatted HTML.
Here you can find the complete text in simple HTML format.
Debian Internal Projects
Debian Med
Free Software in Health Care
Debian-Junior
Free Operating System for Children
(page 1)
Overview
- Motivation for Debian internal projects
- Status of free software in special fields
- Why is Debian suitable for those projects?
- Technology: meta packages, menu system
- Why not just forking from Debian?
- Summary: goals of Debian internal projects
- Future
(page 2)
Motivation
- Profile of target users:
- less technical competence
- not able to install upstream programs with acceptable effort
- no interest in administration
- interest in defined subset of available free software
- need for easy usage
- defined security profile
- Other possible projects:
- multimedia
- authorities
- office
- enterprise / accounting
- education
- ???
(page 3)
Status of specialized free software
- Orphaned upstream
- Strange Licenses
- Hard to use because of lacking GUI
- Porting issues (architectures, byte order, 32 vis 64 bit)
- Data format not exchangeable
- Parallel development of the same functionality
- Used tools or back ends are not intended for those applications
(page 4)
Examples of free medical software Medical Practice
- GnuMed
→ robust, secure, portable, professional design (PostgreSQL+Python) not ready yet - FreePM
→ no professional database back end (Zope), already in production - FreeMed
→ no appropriate database back end (PHP + MySQL), development stalled - Tk Family Practice
→ self made database back end, portable (Tcl/Tk), already in production - OIO - Open Infrastructure for Outcomes
→ out-patient care, robust, flexible (Zope+PostgreSQL), only browser front end - SQL Clinic
→ psychiatric housing, PostgreSQL + Perl web interface - ... and others ...
(page 5)
Why
does the world need this
whole lot of
free
programs to manage
patient data
? ?
(page 6)
Reasons for variety of patient systems?
- historical reasons
- less popularity
- different technological decisions
- different Concepts
Goal of Debian Med: Integration
(page 7)
Example: Dental Practice
- OdontoLinux
→ PostgreSQL, PHP, Debian package available - LinuDent
→ Tcl/Tk, shared code with Tk Family Practice
Authors now are in touch ...
(page 8)
Microbiology
- Debian contains a number of packages to analyze DNA and protein sequences
- different state of development
- different licenses
- sometimes not DFSG compatible
- avoids wide distribution
- problems
- incompatible data formats
- problems on different hardware platforms
- different usage
Goal of Debian Med: Apply Debian standards of quality
(page 9)
Missing free software solutions
- drug database
- pharmacy
- physiotherapy
- veterinary practice
Goal of Debian Med: Supporting developers
(page 10)
Special fields
- medical imaging
- different proprietary image formats of medical devices
- licensing and patent problems
- control of medical devices
- replacements of proprietary software on non-free systems with drawbacks regarding to functionality and security
Goal of Debian Med: Indicate advantages of free software to programmers
(page 11)
Complex problems
- medical research
- powerful Applications (DHCP / VISTA)
- telepathology (ipath)
- medical record
- GEHR (OpenEHR)
- OIO
Goal of Debian Med: Demanding solid packaging
(page 12)
Documentation
- "traditional" weakness of free software
- Possibility to give contribution without high technical skills!
- Medicine HOWTO
- part of Linux Documentation Project (LDP)
- description of existing free medical software
- translation would be great
- Resmedicinae Analysis Document
- analysis of existing programs
- specification of requirements on practice management software
- only rudimentary English translation
- need for this kind of analysis documents for other fields
Goal of Debian Med: Packaging documentation and translations
(page 13)
Debian-Junior
- Goal: make Debian an OS that children of all ages will
want to use
- making it so easy that children are not afraid of
- making it so nice that they like it
- making it a playground for children experiments
- not only providing games but care for their quality
- Debian-Junior is the first internal Debian project
- Debian Med adopted ideas and techniques
Basic idea: Do not make a separate distribution but make Debian fit for special purpose instead
(page 14)
Why using Debian for this purposes?
- powerful packaging tools
- strong quality assurance
- developed by about 1000 volunteers
- support of 11 hardware architectures (auto builders: alpha, arm, hppa, i386, ia64, m68k, mips, mipsel, powerpc, s390, sparc)
- single developers have influence on development - they just have to do it
→ Sub Projects
(page 15)
Why Debian as platform for medicine
- stable, solid, carefully tested (BTS)
→ security - strict rules (policy)
→ reliable, clearly defined installation - variety of hardware architectures
→ universal and scalable - ability to have influence
→ independence from commercial distributor
(page 16)
Technology (1): Meta Packages
- providing meta packages
- defining dependencies
- adapted configuration for certain tasks
- collection of field specific software
- no research for available solutions necessary
- easy comparison
- documentation
- packaging and creating of relevant documentation
- translation
(page 17)
Advantages of meta packages (1)
- defining dependencies from packages for certain tasks
- user is not forced to browse the whole package list of Debian
- easy installation of specific software
- no accidental remove of necessary package, because pinned by meta package
- conflicts to incompatible packages
- apt cares for the proper installation of dependencies
- easy installation
- guarantees stable system
- low effort for administration
(page 18)
Advantages of meta packages (2)
- added value for general Debian user
- Debian policy does not forbid to install junior-games-card even if you work in an office :-)
- installing junior-coloring provides you with a desktop of easy to use painting tools
- med packages for medical office tasks (not yet packaged) could be used in any other office
Demonstration at the end of this talk.
(page 19)
Technology (2): User menus
- define a list of sub project users using debconf to manage groups like junior and med in /etc/group
- provide those users with a menu structure for their tasks
- do not bother other users with those additional menus
- provide script update-subproject to update user menus after new package installation or additions of users to those groups
- not implemented yet
- current Debian Med packages give an idea of the look and feel
(page 20)
Why not just forking from Debian?
- separate distribution would cost extra effort
- would be hardly better than Debian
- continuous work to keep base system boot-floppies etc up to date
- nearly impossible to get security fixes as fast as Debian
→ Forking would be a bad idea
- integration into Debian has advantages
- huge user base all over the world thus becoming public on the back of Debian
- secure and stable system without extra effort
- Bug Tracking System for free
- infra structure (HTTP-, FTP-, Mailserver, PKI, ...) for free
- bringing back something to Debian
- enhancing quality of packages by making them more user friendly
(page 21)
Summary: Goals of Debian Med
- integration of medical software
- apply Debian standards of quality to medical software
- supporting developers
- indicate advantages of free software to programmers
- solid packaging for huge software which is hard to install
- providing documentation and translation
(page 22)
Future
Debian-Junior- first version will be released with woody
- included in task selection list of install CD
- only first steps
- some meta packages ready
- great interest from developers and users
- first version in Woody+1
- Knoppix based Live-CD
(page 23)
Further Information
(page 24)