Software models and architecture
Aeronautics was one of the earliest, and still is one of the foremost, consumers of critical and complex interactive software. Beyond pure design issues, the production of safe and highly usable software for pilots and air traffic controllers raises challenges of its own. One is software verification; no satisfactory solution for specifying and verifying interactive software is yet available. Another is software engineering; with traditional architectures and development processes, design and project management often have conflicting schedules. We believe that the solution to these issues lies in the identification of appropriate models, architectures, and ultimately languages to describe interactive software. Adequate production and verification methods will derive from these.
The Interactive Computing Laboratory explores several directions toward such models and architectures, all rooted in the strong conviction that interactive software is intrinsically not the same thing as pure computing software.
The I* model
Related projects
Istar, ShareIT, Medusa D3CoSRelated publications
Linux input
Researchers had suggested it for decades: humans do not have a left hand with 101 fingers and a right hand with 1 to 3 fingers. Now that the computing industry is realising this, we need operating systems with a rich enough input model. Contributing to Linux-based operating systems is a good reality check for the lab's research on input management in interactive software. It can also help the aeronautical industry to better understand some issues associated to modern input systems.
The lab's involvement in Linux started with multitouch devices. Ultimately, we hope to be able to contribute to the evolution of input management so as to support a wider range of input configurations, and even multimodal input. Read more