Accessibilité Contact Aller au menu Aller au texte

Industrialization & agility : the HELIOS project

Intervenant(s) : Laurent Laudinet
Date : Mercredi 7 juillet 2010 Horaire : 10h00 Durée : 40 minutes
Lieu : ENSEIRB - Amphi C

The presentation

PDF - 4 Mo

The heart of industrialization is traceability, change management and qualification. Such activities are strongly tied to heavy processes and may seems in total contradiction with modern agile development. But if we give it a second though, agile development processes require a greater monitoring of projects, in terms of release control, dependencies identification, and so on… This is very close to industrialization concerns. With proper tools this is a great opportunity for both aspect of development to converge and reach maximum efficiency with a minimum effort. Fortunately, the Open Source communities provide tools covering most if not all the needs.

Modern tools provide facilities for such concerns. The Maven2 representation of a project is a great way to manage its dependencies, providing a fine grained control over the releases of a project. Continuous integration is a powerful tool for developers. When we apply it to the overall project we are creating a new form of change management. This may be extended to many tool that are currently used in the Open Source world. From test campaign management systems to bug trackers, all the commonly used tools are covering the entire field of industrialization and a new way to achieve industrialization is emerging from the use of these tools.

The next step is Application Lifecycle Management (ALM) systems. Such systems is creating the missing link between all these tools and add the missing temporal dimension. This is the purpose of the HELIOS project. This project is Open Source. The HELIOS platform aims at providing reports from the various tool used in a industrialization/quality process. With HELIOS we want to promote this new way to industrialization by the use of Open Source tools and technologies. In an agile development environment, tools are very important. They are heterogeneous, some are used by the whole community, some are used by few. No tools impose a common process as everyone is using the best tool for its own purpose. So in order for this ALM platform to reach maximum efficiency in an agile development environment it must be non intrusive. This implies that one cannot enforce a complete engineering model because it cannot match the multiple models from the many tools used. The issue for the HELIOS platform is to create link between unknown artifacts. This can be solved using semantic technologies. If engineering artifacts are expressed from the various tools with XML/RDF, it is very simple to create and maintain a link between them. Of course such solution have a drawback ; as it is impossible to know what artifacts to bind, it is up to the user to create meaningful links.

In a way, the HELIOS platform provides reports from engineering tools and a way to create semantic bonds between engineering artifacts. It is possible to relate this activity to another IT field : BI (Business Intelligence). BI tools are creating reports out of complex data and provide interesting and flexible methodology. By using Open Source reports engine the HELIOS contributors were able to easily and efficiently create reports from engineering data. Finally, the HELIOS platform can be seen as an engineering system complementary to IDE (Integrated Development Environment) and Forge. It is not dedicated to producing source code as an IDE, nor it is dedicated to collaborative work as a Forge. HELIOS rely onto the Bug Tracking system to integrate with the classical organization of development infrastructure (IDE + Forge). Its human-machine interface is web based so it can be easily integrated. Moreover, a portal solution to access engineering data may contribute to trust and transparency of IT projects by providing customization of audit facilities for final users.

Through the HELIOS project, we all hope to contribute to an innovative use of engineering data.

The author

Laurent Laudinet graduated from the ESME Sudria, a french engineering school (Paris), and holds a Master Of Science in client server computing at the San Jose State University (California). After a short experience in the service industry he has joined the THALES group first as a developer on projects concerning critical defense systems both for national and export clients. Then as a software engineer he was in charge of heterogeneous multi-disciplinary components as well as NATO projects components in an international team. After 4 years he moved to the Open Source Software expertise centre of the THALES group as a software architect.