Category: PFTCR

LD2SD, Helios and Linked Data / SemWeb extracted from development forges

We have received a visitor from DERI last week (PhD student Aftab Iqbal) who’s researching integration of facts about software development tools into Linked Data in order to provide interesting “semantic mashups” of data into IDEs like Eclipse (see his slides). Quite interesting is the choice of ontologies and the results integrated in an Eclipse plugin made available to developers.

This approach is quite similar to the one we practice in the core of the Helios platform (still under development) to integrate data coming from different FLOSS ALM tools in order to create dashboards offering a consolidated view of software (maintenance) process.
Maybe the difference is that Helios does this internally inside a “self-contained” platform whereas the potential of LD2SD presented by Aftab is to do the same on the Web of Linked Data.

Also, in Helios, there are other contributions made for the Mandriva distribution (with links with projects like Scribo and Nepomuk to which Mandriva is also participating) in the form of the doc4.mandriva.org, in order to aggregate, this time, not facts at the “project” level, but for a meta-project (a GNU/Linux distribution) that are quite interesting. See Stéphane Laurière’s slides for details.

We’re also experimenting in the frame of the COCLICO project on producing RDFa data about software development projects hosted in FusionForge instances. (see our progress tracked through this FusionForge feature request). First candidates are project’s DOAP profiles and developer’s FOAF ones, and lots of SIOC to glue it all, and of course other informations relating to a Forge ontology that we’re proposing in COCLICO.

With recent announcements that Mylyn is investing a lot in OSLC, and OSLC being based on RDF, and the advent of the Semantic Desktop starting to emerge (in KDE mainly) on top of Nepomuk, this brings great promises for a great Semantic future.

New SF.net project for HELIOS

The Helios project is gradually going more open, as we start releasing and committing in the open into a SF.net project (heliosplatform).

Among the tools offered by SF.net we will use a blog (wordpress), the wiki (mediawiki) and the SVN, for a start.

The project’s SVN repo will be populated with all components we have developed, as we progressively switch our SVN hosting. The first piece we have committed is the Mantis OSLC REST server module.

Some news of our efforts around OSLC-CM and future plans

OSLC-CM V1 is a proposed standard for REST APIs of bugtrackers, and in our seek for more interoperability in the bugtracker space, we’ve been very interested in it.

OSLC-CM is quite young and only so far implemented in proprietary tools (although elaborated in an open way) on the server side, and as we believe in FLOSS, we’ve started trying to implement basics of server side plugins for a few bugtrackers.
In addition to a demo server that’s simulating the behaviour of a bugtracker, we have started implementing a Mantis plugin and FusionForge and Codendi trackers add-ons (all PHP and based on Zend framework, see this project on picoforge). All are very basic, but we hope they will be the basis for future OSLC-CM compatible servers in these tools.

At the same time we’ve been experimenting with the code already published in Mylyn to support OSLC-CM on the client side. Not everything is public yet in Mylyn, as the elements that have been developped for some connectors of Tasktop to the proprietary tools are being ported to the open source code of Mylyn.
We have thus been able to use the Junit tests classes of Mylyn and tweak them in a way to connect to an instance of the demo server for Mantis (including handling some Basic auth), and be able to retrieve the first bugs descriptions 🙂

Now that this works, we’ll try and add some Java code (maybe reusing Mylyn client libs) to doc4 (being developped as part of Helios) in order to start linking doc4 and Mantis so that this can be used in the Helios platform. This may involve mixing code of XWiki and Mylyn… hmmm… well, we’ll see.

Next steps may be also to try and implement a connector in Python that might be used in tools like bts-link.

Then whichever Python or Java client libraries we have, will allow us to use them inside FetchBugs4.me to connect and harvest bugs of OSLC-CM compliant bugtrackers eventually.

Lots of interesting developments ahead. Stay tuned.

COCLICO started : many interesting development in forges ahead of us in the 2 coming years

We have started the COCLICO project this friday, with a meeting grouping many actors coming from various french regions, that operate in the area of open source forges (around FusionForge, NovaForge, Codendi, Trac, PicoForge, etc.). It’s a “Pôle de Compétitivité” (french R&D clusters) project which is funded by french public agencies, under the frame of both the FLOSS thematic group of System@tic (Paris) and Minalogic (Grenoble).

COCLICO will last 2 years and will let us all collaborate on producing FLOSS components that should allow much more interoperability between the open source forges, and probably deliver interesting standards that should allow to integrate forges with more tools in order to support new uses. We have no website yet, but it will be setup next week.

Of course a collaboration project with many companies (with various profiles, from the single consultant to the very large corporations) and academics is always requiring some effort so that everyone collaborates, but we have a strong focus on producing code as first steps, and I’m quite confident we all believe that FLOSS is necessary to share the innovation efforts.

I hope it will be a great occasion to bring interesting new things in the FLOSS ecosystem, and that we’ll manage to let others participate even if they are not funded by COCLICO, since one of the goals of the project is to bring momentum in the general forges ecosystem.

As far as we’re concerned at Institut TELECOM, we’re leading two workpackages on interoperability and community/ecosystem.

I’m very excited about this project, which together with our running Helios project should allow us to contribute in a significant way to FLOSS development tools and to the general quality of the FLOSS development process.

Expect more spamming from me about forges in the future on this blog 😉

Update : we now have a website both with more details in french (including a description of the project’s work-packages) and in english (still empty at the moment, working on it).

First webcast of a demonstrator of our bug ontology’s use

We have setup, as part of our work in Helios, a very early demonstrator of a database of RDF facts about bugs in several distributions (currently Debian and Mandriva), in order to try and validate the Ontology describing bugs that we develop.

Here’s a pointer to the first webcast on fetchbugs4.me’s blog, with more details.