The EU now provides a search tool to find applications among the 1751 open source development projects hosted on ten federated forges managed by Austrian, French, Italian and Spanish public administrations.
The new search engine basically relies on an automatic translation service, translating projects’ descriptions in English.
The Open Source Initiative board, after visiting the European Commission, has finally approved the European Union Public license on the 4th of March.
The EUPL 1.1 – the revisited version of the EUPL 1.0 including recommended modifications resulting from the OSI discussion – is supported by the EUPL community. Stakeholders can share opinions and pose questions through the EUPL forums and blog.
Hi Roberto,
Publishing legal instruments with equal value in all EU languages and letting European courts work with it, is really “business as usual” for the European Institutions: Community Law counts million pages, valid in 22 or 23 languages. In case of uncertainty, national courts may ask the European Court of Justice to clarify. It is a simple fact that the GPL V2 and V3 (relative) failure is that FSF was unable to deal with linguistic diversity (that they consider as a major risk; it is true that they do not benefit from a supra-national court to help). Having said that, the EUPL has other merits: the license complies with European flavour of copyright, information to consumer, warranty and liability. The most important however, is that the intention of EUPL promoters is not to compete with the GPL or any other license. The objective is long term: to bring more administrations to license their source. We are at the early beginning of this, but there are currently substantial move in this direction and several Member States consider policies to license under EUPL. Last, the EUPL has a compatibility list (with GPLV2 and other copyleft licenses). This is also unique: it would be interesting to try solving license proliferation through “mutual recognition” between a club of equivalent licenses, sharing the same compatibility list: this would also provide more freedom to developers. Who has comment on this idea? It is a fact – unfortunately – that “compatibility” is seen by the FSF as “compatibility in the one-way direction of GPL, GPLv3 and AGPLv3”. The reverse situation is not even imagined by the FSF, (and by the way, their current web site ignores totally the EUPL – remember Mahatma Gandhi: “First they ignored us…”). A last word about the “biased” comparison I did between the EUPL and GPLv3 during the EOLE event in Paris: Of course you can say it was biased! (since I am a EUPL advocate), but please do not take my comparison as an attack “against” the GPLv3: I just report facts: GPLv3 is about 3 times longer, is full of technical details, is complex (even for a lawyer) and as it is officially valid in English only it may not be the most persuasive license for a German or French administration. Could anyone object? Now if you are happy with GPLv3 and AGPLv3 no problems, go on!
Hi Patrice-Emmanuel,
I am glad you joined the conversation. I think that having a license (EUPL or not) translated in all European languages is a great thing.
What I object to the EC is that you decided to go your way, instead of participating the GPLv3 process, managed with a public and transparent consultation.
What I object to you is your adversarial approach. I believe that opening to FSF licenses is your concern, since you are trying to convince European public administrations to use the EUPL.
Show them that is convenient, that they can still stand on the shoulders of giants, and try harder to consider GPL, and software licensed under the GPL, as your best friend, not a foe.
My two European cents
Roberto Galoppini and 
P-E Schmitz are discussing.
Jesus M. González Barahona 6:20 pm on August 31, 2009 Permalink
Well, I’m not that sure about your proposal of looking for *code*. In the case of OSOR, the idea seems to be to link with other sites devoted to libre software for public administrations. Working at the package (or poroject) level, seems to me about right. When you’re looking for some piece of code, the domain of the program probably doesn’t matter that much…
In other words, if you’re looking for some package which may be useful for public administrations, looking at OSOR and federated forges seems reasonable. But if you are looking for a specific piece of code (even if it is for a software to be used in PA), the search should be much wider: probably almost any forge could have the piece you want).
WRT FLOSSMetrics, indeed we’re focused on GForge-like forges (including SourceForge), but we can extract data from any public repository in any forge, provided we have its url, and the kind of repository is supported by our tools. Currently that amounts to CVS, Subversion, git and (limited) Bazaar for SCM, Bugzilla and SourceForge for bug reporting systems, and mbox for mailing lists repositories. More are supposed to come.
Yes, I fully agree that the services provided by FLOSSMetrics could be integrated with OSOR, or with any other forge, for that matter. SourceForge is starting that way (not with FLOSSMetrics, but with their own machinery), and OSOR also started it, offering graphs about the evolution of some parameters related to the activity of the projects (in this case, using a part of the FLOSSMetrics toolchain).
Just to finish, thanks a lot for reporting on FLOSSMetrics, and for taking the time to understand it!
[Disclamer: I’m coordinator of the FLOSSMetrics project, and also involved in the OSOR as a member of the consortium maintaining it]
Roberto Galoppini 7:04 pm on August 31, 2009 Permalink
Hi Jesus,
glad to see you joining the conversation.
As I wrote in my blog post most of the times open source projects for public administrations are lead by SMEs thinking and acting locally. Translating projects’ descriptions can hardly help the share and reuse of knowledge in the context of IT, I am afraid.
Krugle code search engine or similar technology might help to search pieces of code that perform more specific tasks, and eventually reuse code made available from other EU public administrations under the EUPL (apparently designed to ease the licensing burden).
Taking advantage only of code hosted on federated forges may result in a lack of opportunity anyway, either if you look for a whole package or a library. In other words, I am assuming that we need of OSOR here, and the EUPL license may well be the reason for that.
I wish to report more about FLOSSMetrics, let’s keep in touch for writing a specific blog post on project’s final findings.