Contents
Editorial
7i program
Tribune
Hot topics
Business cases
Experts voice
Quick Poll
At a glance
Events
PDF version
 

Subscribe to
Bull Direct:


 

Archives
April 2007
Guest contributors
Bull's service centers
Thierry Lapresle, Director of Bull Service Operations in France
"Control and industrialization are keys to a successful project"
Captain Georges Deli
“Control and industrialization are keys to a successful project”
Interview with Captain Georges Deli
Commander of the French Navy Human Resources Information Processing Center (CTI-RH), and Director of the HR Decision Support Information System modernization project.


Based in the French city of Toulon, the CTI-RH manages all the HR information systems for the French Navy. These systems are used by the 49,000 active service personnel throughout France, and across all the Navy’s military bases worldwide.

Why did you recently launched an ‘industrial’ development project using Open Source?
Every organization is constantly faced with the challenge of delivering the services it has committed to deliver in increasingly short timescales. That poses the problem of how the organization adapts to change. At the CTI-RH, HR reforms are being driven forward as a result of the move to the euro, the passing of the new LOLF financial act in France, and the new status of military personnel. These rapid and permanent changes are challenging the stability of our application that we must absolutely manage. Even if the Navy’s HR information system offered all the functionality that its users required, the user interface for the HR decision support system (known as SIAD/RH) currently runs on an obsolete product and requires a ‘fat client’, which makes it hard to distribute software more widely and to further evolve the system. So, we thought it was time to change the system interface and go ahead with modernizing the architecture to make it more open and accessible to a much larger number of people. Having compared the various possible solutions – ERP packages, .Net, J2EE – and having also compared the overall costs of ownership, we finally opted for a customized J2EE development, using Open Source and Web architectures. Our decision was underpinned by other advantages, such as the greater independence (avoiding ‘captive’ markets is a particularly sensitive issue for us), the skills required (to use the same technologies as the university sector), and cost control. So we decided only to change the front and middle office, retaining the same common database. This approach meant we could develop the new system in a gradual, modular way.

Why was the industrial approach to development work so important in the context of this project?
It was vitally important to maintain a fully operational system throughout the life of the development. There was no question of putting HR reforms on hold, and inevitably the system had to be aligned with the work related to the introduction of new regulations. The existing SIAD/RH system had to carry on evolving until it had been caught up by the SI@D/RH, which was being gradually deployed. So the two systems effectively had to be run in parallel. It was unrealistic to maintain two technical teams, using very different technologies, and while tackling all the other challenges. We needed professional support to help us with this modernization. We chose Bull as the prime contractor, because of its proposal to provide a package of ‘industrial components’ from the Open Source world, and its expertise as a contributor. Bull does a lot more than just provide us with software downloaded from Open Source communities on the Internet: they also interface them so that we are not exposed to complexity. They qualify products and integrate them into our work environment, carry out performance tests, give us advice, and provide training, management and expertise; all with the aim of making us more self-sufficient. To achieve this, we have implemented Bull’s NovaForge Open Source software development factory, including a portal, development tools (based on Eclipse), and integration and testing tools. All this also integrates with our own software engineering workbench, MEGA. Today, the forty-strong project development team – half of them our own technical people and the other half from Bull – can work together in a professional and structured way, with a high degree of synergy and complementing each other’s activities.

What are the best practices to adopt in this area?
No matter what the project may be, all change has to be supported by a strong managerial commitment. This support will be stronger if the decision-makers have a very good view of the project and good tools for managing its progress. Initially one might think that Open Source is not well placed to offer this. It’s a fear of risk that often leads people to choose off-the-shelf products such as ERP packages. However, our experience in the field shows just the opposite: by enabling rapid, staged construction of modular, robust and re-usable solutions, and by offering the ability to go back, and avoiding ‘big bang’ approaches, Open Source enables you to save time, helps you achieve quick wins, gives you more flexibility, and ensures you can give decision-makers the greater visibility and ability to control the project which they want. And finally, we must not forget the issue of support. In effect, Open Source software is, above all, software: subject to the same kinds of risks and success as any solution from an ISV. A contractual commitment with a professional provider of support services seems like the better option as far as I am concerned.

 
Contact  |  Site map  |  Legal  |  Privacy