The Asset Description Metadata Schema for Software, ADMS.SW, specifies a standardised way for description metadata for software projects. The development of the specification is an initiative from the European Commission and brings together a wide range of experts in a Working Group. One of the Working Group members, Olivier Berger, was interviewed on the benefits of ADMS.SW and on how the specification can be implemented on existing software forges.
We have been researching issues with interoperability of development tools use in the FLOSS communities for some years now in our team at Telecom SudParis. In the last years, we participated in the COCLICO project, funded by public agencies in France, which grouped companies and research labs interested by interoperability of software development forges. We investigated the use of Semantic Web technologies to foster interoperability between the forge tools communicating on the Internet during COCLICO, resulting in the publication of a first draft of a ForgeOntology.
Naturally, when the same approach was investigated in ADMS.SW, and after COCLICO had come to completion already, it was a good opportunity to transfer some of the lessons learned and improve our results, with a larger scope, this time at an international scale.
What do you see as being the major benefits of ADMS.SW?
ADMS.SW paves the way to more interoperable software catalogues, using standard technology like RDF, which should help users to more easily discover useful FLOSS software available for their needs.
In a longer perspective, this improved semantic interoperability will foster novel software maintenance practice like Linked Data driven Software Development which should help maintainers of FLOSS distribution, vendors and integrators, to more efficiently interlink quality assurance tools that need to track versions of software, bug reports, and other artefacts which are already available on the web, but that, so far, could only be monitored manually, or using ad-hoc adapters.
What could be potential barriers to the adoption of the specification? How can these be mitigated?
Software development platforms (forges) or software catalogues are often operated with free software components, and the development teams for these components will need to find some motivation to get interested in adopting new specs, new formats, and there's no guarantee that external factors can increase their interest.
Adding to this problem, forges are often maintained by small teams with tiny budgets, in fragmented communities. Reaching standardization on some interchange formats for meta-data description requires some form of consensus, and validation by implementation, which is a long process when a big part of the developers are volunteers with an already too big to-do-list.
Quite a lot of the Public Administration forges in Europe (main targets of ADMS.SW efforts) are using variants of the same software, either old GForge versions or its newer FusionForge fork. If someone motivated enough and able to understand FusionForge's code develops an ADMS.SW compliant feature, and manages to push it in the released versions, there's a hope that sooner or later more and more forges will be interoperable when they are upgrade to a newer version.
You were responsible for implementing a plug-in for FusionForge – allowing any forge’s project metadata to be made available using DOAP and RDF. Will it be difficult to adapt this plug-in to support ADMS.SW?
ADMS.SW reuses some of the meta-data descriptions already popular in DOAP, so naturally, there's hope that the plug-in we developed initially for FusionForge during COCLICO will be extended to become compliant to the novelties brought by ADMS.SW. Still that may require working with the FusionForge project tightly to make sure it integrates in the release agenda, or can be back-ported for earlier versions in case administrators of FusionForge won't be able to upgrade their forges, just to benefit a new plug-in.
What would be the benefits of using an ADMS.SW plug-in?
Any RDF compliant tool will be able to consume the ADMS.SW compliant meta-data published by the FusionForge plug-in. This will offer the projects hosted on the equipped forges the possibility to publish on the Semantic Web (aka Web 3.0) their descriptions in a machine-to-machine format, in addition to the pages meant to be read by people.
FLOSS projects will become participants to the Linked Open Data ecosystem, which is emerging these days, and has only started delivering some interesting promises. It's probably too early to even imagine all the possible uses for the long term, but certainly joins many other efforts in bridging FLOSS and Open Government initiatives.
But in the short term, the goal is to make sure that portals like Joinup will retrieve interesting software projects and categorize them automatically in the business domain relevant to certain citizens or administrations, without too much manual intervention, manual conversions, which are always error-prone, and allow more direct "delivery" of meta-data from the developers to the end-users.
About the proposed ADMS.SW exporter
The ADMS.SW Working Group will ask Olivier Berger to develop an exporter plug-in that supports:
- Exporting metadata at the forge-level, but with forge administrators having control over which projects are exported.
- Exporting metadata at the project-level.
- Exporting to a simple RDF format.
The plug-in will support FusionForge installations of version 5.0 and above.
Olivier is a research engineer in an academic lab at Telecom SudParis, a public higher education school member of Institut Mines-Telecom in France. He is also involved in the development of FusionForge and contributes to the Debian project.