De afgelopen decennia was er grote onzekerheid over de Hubble constante die de uitdijingssnelheid van het heelal geeft: waarden varieerden van ongeveer 50 tot 100 km/sec/Megaparsec. In mei 1998 zijn alle groepen van de wereld die betrokken waren bij metingen met behulp van de Hubble ruimtetelescoop onder leiding van J. Mould (Canberra) twee weken bijeen geweest op het Lorentz Center, om alle data en data-analyse te combineren. Dit was de eerste keer dat de deelnemende groepen in een werkomgeving bijeen waren; elke ochtend om 11 uur vond een centrale meeting in de ``common room'' plaats, waarbij de voortgang en de planning voor de komende dag besproken werd. Aan het eind van deze periode kwam men tot de conclusie dat de juiste waarde km/sec/Mpc was. Er bestaat consensus binnen de astronomie dat het bijeenbrengen van alle betreffende groepen op één plek cruciaal was voor het bepalen van de nu internationaal geaccepteerde waarde.
In de informatica bestaat minder dan in de wis-, natuur- of sterrenkunde ervaring met workshops van het type zoals die op het LC georganiseerd worden. Voor veel deelnemers aan de workshop The Global Internet: Measurement, Modeling and Analysis in 2000 was de interactieve werkwijze en het feit dat deelnemers over een eigen werkkamer beschikten, dan ook een verrassing. In een korte samenvatting van de impact van de workshop enkele maanden later schrijft Ogielski, een van de twee Amerikaanse organisatoren:
Anyway, our Leiden workshop has made an excellent impression in the leading edge Internet research community. As a nice example, at the late 2000 DARPA networking meeting people would refer frequently to the Leiden workshoptalks (they're on the web) without any need to explain to the audience what workshop it was.
Getting in one place good people who measure and analyze the global Internet from quite different perspectives was very productive. From my point of view, I'd say that the workshop definitely had impact in at least two areas:
-- Bringing to the front the realization that global routing protocol instabilities are very strongly coupled to packet losses and traffic fluctuations. As one participant said, most of conventional wisdom about routing convergence is wrong". (by the way, until now, almost all of routing research community has been separate from the traffic dynamics community). These very illuminating discusions are leading to new research directions already.
-- Presenting a thoughtful critique and serious discussions of the pitfalls and possible benefits from the applications of concepts from statistical physics and dynamical systems to networking research. There has been very little feedback from networking engineers to physicists doing such work until the workshop.
Tenslotte een commentaar van een wiskundige. Lenstra, die een eerste workshop over getaltheorie in 1999 op het LC hield, heeft een van zijn Amerikaanse collega's, E. F. Schaefer, naar zijn bevindingen gevraagd. Hij schijft letterlijk:
I attended the two-week long meeting at the Lorentz Center on number theory in April, 1999. This was the most productive meeting I have ever attended. There were two days that were full of talks and on the other days, there were only one or two talks. The organizers had carefully gathered people with similar interests. So the remainder of the time was spent with people interacting in offices, working through mathematics together. Several times each day there was a group of four or five mathematicians working on some problem or hashing out an idea. Other times there were smaller groups. No one wasted time working alone. It was extremely exciting.
I was one of four co-authors of a paper, each of whom were in attendance. We showed the paper to William Stein, another guest. He found two mistakes in the paper and suggested new directions the paper could take. We invited him to collaborate on this paper and that began his productive relationship with several of the co-authors. I started another paper with Michael Stoll (that we are almost finished with), one day right after lunch. We made so much progress that afternoon that we were late for a talk. Not all progress led directly to articles. Sometimes we just worked through a concept that no one knew well and we wanted to understand. With this critical mass of people, we always had success.
The progress we all made was due to a lack of talks and the conducive facilities at the Lorentz Center where four or five could comfortably gather in an office and work at the board and a computer. However, the general lay-out is intimate enough that it was easy to notice when something exciting was happening in another office and join in.
Nils Bruin was finishing his dissertation at the time and was intending to go into industry. Since the organizers invited people with related work, he got to see how he fit into this community of number theorists and what they thought of him. By the end of the workshop, he decided to stay in academia. We are all very pleased.
I also attended the pre-ANTS workshop at the Lorentz center in June, 2000. Again, the organizers carefully chose guests with similar interests. That and the excellent facilities at the Lorentz center produced a valuable experience. This time, unfortunately (and unavoidably), there was a number theory workshop in Utrecht during the same week. So talks filled up much of the time that people were in Leiden. Some people who did not traditionally fit into the clique from the previous year (like Paul Gunnels) were invited. It was worthwhile to create bridges between the group who had met the year before and these outsiders doing related work. With what time remained, there were some interactions in offices like there had been in the previous year.
I would like to add that the administrative staff at the Lorentz Center is extremely professional and a pleasure to work with. I am grateful for the time I have spent there.