|
|
| Front Page | INédit | Contents of n°64 | Conferences and Events | Press |
|
 |
Pierre Paradinas © INRIA / C.Lebedinsky |
: What are the highest-priority objectives for the Department of Technological Development?
Pierre Paradinas: Since the department was created one year ago, we have focused our efforts on setting up structures and resources to facilitate mastering the software development process. Software is one of the unique points of our discipline. It is a research tool used to demonstrate the ideas and concepts proposed by researchers that can be distributed and reused immediately, either for research purposes or to meet the needs of companies. Still, the software needs to be corrected, improved and modified to adapt to the context in which it is used, as well as changes in the software itself and the world of technology in general. A certain level of maturity is also required to allow communities of researchers to capitalize on these results and increase knowledge production more quickly. This is the only way that "in-house" software can hope to be adopted as the software used by a community, or even as a near standard. The question at this point is – who is responsible for bringing the software to life? It definitely does not fall under the domain of basic research, and yet technology transfer is certainly part of the Institute's key tasks.
: What software needs to be developed, and how do you ensure that it is developed?
Pierre Paradinas: The end goal is to ensure that all software is built in a way that makes it easy for third parties to adopt. Indeed, we can never predict whether a piece of software can be usefully deployed for industrial purposes. Someone might find a use for it several years later, once the expertise has disappeared and there is no coherent documentation available. In fact, the experimentation and development departments (SEDs) were designed to support researchers in this effort to develop software (see "Software Crafted with Care," Inédit No. 55). They help researchers to resolve any engineering problems that come up and that will influence the way the software is developed. In the future, these departments will be called upon to play a more important role by working with researchers to support the technological development initiatives (ADTs) that we have recently implemented.
: How is this new tool unique?
Pierre Paradinas: Technological development initiatives (ADTs) are an incentive-based organization system created at the end of 2007. Researchers and engineers in SEDs collaborate to build a technological development program, which is then submitted to the Department of Technological Development. Projects that are chosen receive the resources they need, particularly in terms of engineers assigned from the SED who will participate in the specification, design and execution phases, requiring planning, follow-up, traceability, quality assurance and most importantly skill meshing within the Institute.
It is a structured, large-scale approach: one ADT can involve three to twenty people. It offers an opportunity for researchers working on different projects to meet and propose ambitious programs and initiatives that would not have been possible previously. New directions for research can emerge from this process that might not even have occurred to the originators before they worked together. This prospect is highly motivating for researchers.
The first ADTs will begin in 2008. The call for proposals for these initiatives is scheduled to end in early April, and the number of applications already received demonstrates how enthusiastic researchers are about this approach.
: Do you have any other initiatives in mind to encourage software development?
Pierre Paradinas: One initiative that is already well underway is the effort for software evaluation (which item should be evaluated, based on which criteria, and why). The initiative, led by Daniel Pilaud and the D2T, will involve many players within the Institute as well as external participants.We are also currently looking over our technological platforms, which are essential not only to the visibility of work performed at INRIA but also in our dealings with our most frequent academic and industrial partners. The process has two objectives: first, determining at what point to create a platform or decide that work performed for a certain piece of software has become a platform, and second, determining the optimal mode of governance for the type of structure in question. To do this, we must reconcile the interests of users, who want a stable product, with those of researchers, whose role is to constantly develop the platform.
Pierre Paradinas
Director of Technological Development (D2T)
Tel.: + 33 1 39 63 51 51
