HELIOS project has its domain

We’re busy on the first tasks of the HELIOS project… nothing spectacular to announce publicly, so far.

Still, we have registered a domain for the project (http://www.helios-platform.org/ … which currently redirects to our internal workspace on a LibreSource forge), and there are minimal informations about the project available already here.

I hope we’ll have news to provide soon about the project progress. Keep tuned 😉

P.S.: we’re going to hire an engineer for HELIOS to work with us in Evry. If you want to work on a R&D project involving Free/Libre/Open Source software (bugtrackers infrastructure, etc.), don’t hesitate to get in touch… more detailed position offer to come in the future.

Helios day 1

Just a quick message to announce the start of Helios project, whose kick-off occurred yesterday in THALES premises in Palaiseau.

The team is enthuisastic, I think, and I hope we’ll be able to provide usefull integrated tools for Application Lifecycle Management based on Open source tools.

I hope the work we’ll be doing on bug trackers will help improve the process of bug fixing in open source projects and distributions too.

The project doesn’t have a website yet, but things will be setup soon.

More on Helios later in this blog. Stay tuned.

Introducing project HELIOS

We’ve been setting-up the codename HELIOS project, together with other partners (lead by Thales), which got funded recently, as part of the pôle de compétitivité System@tic (libre software thematics).

That means that our PFTCR team at Institut TELECOM will be able to do more research around infrastructure and process of production for libre software projects.

Among the things we’ll focus in particular in this 2 year project, is the topic of traceability between bugtrackers (WP3). This topic seems quite interesting to many people I’m talking to at the moment in libre software projects (as discussed recently at RMLL/LSM) : good 😉

I hope we’ll be able to help improve the current state of the art where lots of manual synchroniation is done in libre software (between upstream bugtrackers and distributions’ ones, for instance), and deliver useful tool to the communities.

We’ll keep you posted as the project moves on (scheduled kickstart september 2008).

For more details (in french), you may see the brochure (page 4) at System@tic’s site.

Update 2008/09/17 : here’s the official description of the project we’ve just drafted :

HELIOS is a project related to the System@tic research cluster. The aim of the project is to provide an Open Source ALM (Application Lifecycle Management) portal allowing to test, integrate, configure and maintain the many components of an application. HELIOS must be flexible and extensible enough to adapt to any tool with limited developpements in order for the users to keep on using their own tools. In the same way, flexibility and extensibility should allow services providers to build commercial offer. Most of existing ALM tools are specialised into one particular field (requirements management, qualification, project management, …) but the HELIOS project will aim at providing a complete platform covering activities from qualification to maintenance.

Report from LSM/RMLL 2008 on communautary development track

There’s probably much more to say than I’ll remember, but here’s an attempt at reporting from the excellent edition of RMLL/LSM which was held in Mont-de-Marsan (France) early july.

I’ve been chairing one of the tracks, on Communautary development, where I’ve had the pleasure to chair and attend excellent presentations. The rest of the LSM/RMLL was very good too, but being stuck in a room, I couldn’t attend much of it 😉

To summ-up, there have been very interesting talks and discussions on the following subjects (links to descriptions of talks and their slides included) :

  • translations : Claude Paroz has presented the classical process of translation in libre software (gettext, etc.) and organized a practical workshop to help get contributors started. But just before his talk, Marc Laporte (aka the man paying free beers at night) presented a system which was implemented in a wiki to handle multilingual content, which seems very smart, in helping synchronize multilingual content in wikis (where individual translations may change in a non-coordinated way). I think that both talks were very complementary : great to have had both speakers there… and by the way, they proved the international nature of the RMLL (Switzerland and Quebec/Canada) 😉
  • forges : another topic was the forges, or the development environments in (potentially) large projects, with the presentation of the forges genealogy and the GForge project made by Roland Mas. It was interesting to get feedback from the audiance where people reported from their switch to GForge AS, for instance. Also a presentation by Quang-Vu Dang about the use of semantic web standards to monitor activity in forges. We also discussed the semantic web standards and interoperability after other presentations about bug-tracking or packages (more bellow). Lucas Nussbaum also presented the infrastructure of the Debian project which loosely integrates different tools which are used to monitor the activity and do the QA work in the project. Lucas’ presentation was too short unfortunately, for such a complex project in-depth review (and trolls popped-up also ;).
  • packaging : Lucas also presented interesting starting elements for attracting volunteers contributions to Debian, by describing the packaging of applications in Debian (and Ubuntu, sort of ;-). Complementary were the presentations by Vincent Untz and Bruno Cornec, resp. on the OpenSUSE build service, and Project Builder, which both more or less manage the generation of packages for various distributions. Their philosophical approch seem different, which lead to interesting discussions : is upstream supposed to get interested in specifics of package contruction in various distributions, or should it be handled independently ? Great debate. There were also intersting talks about convergence in package description formats, which would need more detailed discussions (I welcome any links).
  • Release process : we had three talks which addressed this topic : first the excellent (and crowded, although very early in the morning : 9:00 😉 ) presentation by Thomas Petazzoni on the Linux project process. Next Lucas’ presentations on Debian (comparing release strategies between Debian and Ubuntu, for instance). And finally Vincent Untz’s other presentation on the 6 month paced release process in Gnome. Very complementary and interesting talks, IMHO.
  • Bug tracking : Of course this was the topic addressed by Emmanuel Seyman in his very interesting talk about Bugzilla. But we also discussed the subject of bug trackers in Lucas’ presentation, for instance (with the Debian BTS), or when we discussed the problem of synchronisation of the bugs lists between upstream and distributions (which will be one of the topics of our forecoming HELIOS project : more blogging ahead). Definitely something where the contacts were very valuable amond people attending and presenting.
  • Other topics : well, that wasn’t all with this track at LSM/RMLL, but I wasn’t as much interested in these others I guess. You’ll find more details on the conference’s site.

I hope the content was enjoyable to the audiance too (although I disturbed the presentations with my silly jokes or my facist approach to schedules ;).

See you in next edition.

Codex liberated: here comes Codendi

Codex was the name of the forge developped by Xerox on the code base of SourceForge. Although Xerox once contributed parts as free software which ended-up in libre versions of the sourceforge codebase (in Savane or GForge, if I remember well ?), it seems there were still parts that were proprietary (I’m not a user myself, just talked to people who know better).

Although it was used by numerous projects at clients of Xerox (and internally), I’m not sure the strategy was clear for Xerox, mainly considering the cost of maintenance of these complex platforms. Xerox made a business of service around Codex, but wasn’t so sure about the licensing strategy, I think, and wether being an editor for Codex was something really interesting. There were some trolls around when the subject was mentioned, together with the interesting features in Codex that lacked in GForge, for instance, when would they contribute to the free world, etc.

They have apparently make up their minds and liberated the parts of the code that weren’t free, and also renamed the software which is now called Codendi (it had passed under my radar, but I got the news at the workshop of french forgerons).

This opens new opportunities for collaboration for Xerox and the users of Codex/Codendi, and may help achieve some convergence between GForge and Codendi, on parts which keep quite similar in both software which initially started out of the last free versions of SourceForge (through plugin infrastructure unification ?).

After the liberation of Scilab, yet another new libre software !

Let’s hope that this is a very good news for the future of the forges in general.

Congratulations Xerox !