The European Commission sent out a
on July 8 announcing a new report it had published
on sharing of open source software between European governments. Those who
are interested can get the full report as a 3 MB
; for the rest of you, we have read it through and distilled
out the main points.
The focus of this report is relatively narrow. The Commission is not
trying to promote open source in general, and it is not trying to get
governments to use free software desktops. Instead:
The object of the study is the specialized software produced by the
public authorities across Europe, to respond to the administration
or more generally to eGovernment needs: administration of roads,
hospitals and public health, education, tax payment and recovery,
justice, territory management.
In other words, the Commission thinks that open software can maybe help
keep the trains running on time. Performing these governmental tasks
requires large amounts of custom software. In general, there is not a
market for this kind of administrative software, since there are so few
buyers. So governments end up writing their own. And that, of course,
leads to the obvious question:
...rather than to develop nearly identical solutions separately,
why not adopt the open source development model to share the cost
between a broader (trans-border) development team?
The open source case is helped by the fact that governments will, in
general, need to be able to adapt any shared software to their particular
So the report's authors envision setting up a "Pool of Open Source
Software" (POSS) portal where governments could share their software. The
end result looks very much like a multilingual, restricted access version
They have already picked out the components they expect to use in the
creation of this portal: Linux, Apache, ProFTPd, MySQL, phpMyAdmin, exim or
sendmail, mailman, python, fetchmail, webalizer, PHP, cvs, sourceforge,
OpenSSH, etc. They picked
open source tools "to reinforce the credibility" of the project, "although
we do not consider this requirement as a technical one." Running this
project is expected to cost about EUR 6 million over five years.
Much space is dedicated to worrying about licenses, patents, and
liability. Governments, it is said, satisfy two criteria that make them
especially prone to litigation: they are easy to find, and they have deep
pockets. So a licensing or liability issue that attracts little attention
when a small company or development group is involved could turn into a big
court case for a governmental agency. To avoid such troubles, the report authors
want to nail down a number of legal items with more than the usual amount
For example, very few free software licenses specify where any disputes
should be resolved. The report states that the license for any software
distributed through POSS should be augmented (with a separate agreement,
perhaps) by a statement of jurisdiction. If a licensing issue goes to
court, they want to know which court. Similarly, they want a
declaration of which country's laws apply in a dispute.
Patents are a concern as well; the report seems to accept that software
patents are in Europe's future. There is a discussion of an IBM submarine
patent in the ebXML specification as an
example of the sort of trouble that can come up. The report concludes:
A practical consequence of software patentability regarding the
publication or the pooling of open source software inside the POSS
is the requirement to investigate on possible patents, in order to
avoid legal hassles and even higher costs.
The report has no suggestions, though, on how to find all of the
potential patent problems in a given piece of software.
Then, there is the issue of liability for software-related problems. The
report writers worry that the standard liability exclusions found in both
free and proprietary licenses may not be legally valid. They hope to
address this problem by instituting a review process within the POSS system
- though it's hard to imagine how this group could, with confidence, issue
a clean bill of health for any package.
There is one other component to the report's solution to licensing, patent,
and liability issues: restricting access to the software to "public
administrations," initially in Europe only. With a restricted user base,
contracts can be signed that give the POSS system - and those contributing
software to it - a better handle on the various legal issues. A "public
administration" which obtained software from the system could, of course,
redistribute it under the terms of its (open source) license; they would,
then, take on the related legal issues. In practice, it would not be
surprising if very few government agencies redistributed software obtained
In other words, the software involved may be open source, but there are
limits to the openness of POSS. European Union citizens wanting to look
at the code used by their government may have a hard time getting access -
even though said code is, in theory, under an open source license. POSS
looks more like a private code sharing club than a true open source
project. Sharing code may be helpful for governmental efficiency, but the
"members only" approach could deprive both governments and citizens of many
of the advantages of truly free software.
Comments (7 posted)
The Electronic Frontier Foundation has issued a
on the abandonment of the 2600 DVD case, which will not
be appealed to the Supreme Court. This marks the end of one of the more
prominent DMCA cases, and it sets some unfortunate precedents - at least,
in the second federal court circuit. The ban on a piece of software as a
"circumvention device" remains intact, and, chillingly, it is fine for the
government to prohibit linking to content that it does not like.
The EFF's position is that this is not the right case to take to the
Supreme Court - the end result would be much the same as with the lower
courts. It is true that the EFF's resources are limited and should not be
expended tilting at windmills. One can only hope that the right case comes
along and we can begin to put a stop to the erosion of freedom in the name
of protecting intellectual property.
Comments (3 posted)
A few small changes have been made to the site, in response to user
requests. They include:
- There is a new combined security page with a
snapshot view of the various security-related resources on LWN. If
all goes well, we'll put together similar pages for other categories
- The weekly archives page is back - and
includes the table of contents for weekly editions published on the
- Comments are now presented in full-text form after an article by
default. This behavior has always been available to readers with
accounts. (If you do have an account, we assume that you set the
option the way you wanted it and thus we did not change it).
There are many other enhancements we would like to make to the site if we
can keep things going long enough. If you have not already done so, please
consider donating to LWN or
advertising to help keep LWN on
Comments (3 posted)
Page editor: Jonathan Corbet
Inside this week's LWN.net Weekly Edition
- Security: Squid-2.4.STABLE7; Linux immune to bind vulnerability
- Kernel: The end of kiobufs; 2.5 IDE troubles; How scalable is too much?
- Distributions: News from Debian, Mandrake; Four new distributions.
- Development: gEDA updates, miltrassassin, OpenEMR, AFPL Ghostscript 7.21, ASPseek v.1.2.9,
Zope-CMF-1.3-beta2, Equinox Desktop Environment 1.0-beta, GARNOME 0.12.1.
- Commerce: W3C Patent Policy: Latest News; Linux for Astronomy V7,8,9; Larry Ellison to talk at LinuxWorld.
- Press: EC advises open source, copyright issues, making a living
with Linux, Itanium 2 support.
- Announcements: Ardour Howto, The Perl Review, KDE report for Linux@work 2002,
Ruby Conference 2002, Gallery of Geeks, comp.lang.php.
- Letters: LCA miniconferences; ssh fiasco