Experimental Algorithmics

This seminar is the sequel to the very successful first Dagstuhl Seminar on Experimental Algorithmic s in September 2000 which brought together researchers from both worlds of algorithmics, theoreticians and practitioners. The main question of that seminar was whether and how theoretical and experimental research can co-exist as equal partners under the big roof of algorithmics . Far from coming close to an answer to that question it was decided that there should be another seminar two years later.

Thus, this time we will continue to discuss the fundamental question of the value of experiments as opposed to purely theoretical analysis of algorithms. But this time we will also ask what happens if computer scientists (theoretical or practical) venture out in the world of real systems building and testing (networks, bioinformatics, natural language systems,...) where they usually meet non-CS engineers or physicists with their own methodological framework of experimental evaluation. Is there a fruitful interaction between CS and non-CS? Can we (the experimental algorithmicists) learn from them? Or they from us?

The aim of this workshop is to bring together three groups, theoretical oriented researchers, practical oriented researchers, and people working on real systems. We will invite a few keynote speakers for longer talks, but otherwise the seminar will follow the usual Dagstuhl tradition of short talks and much time for discussions.

The seminar will be held in the week before the ESA 2002 conference in Rome (ESA 2001: http://www.brics.dk/esa2001/)