Date: 10 Sep 2004 From: jyrki@diku.dk Subject: Re: FIRST news Gitte: Here you get a corrected version of my text that you can publish on your webpage. \Jyrki ======= In connection with the Scandinavian algorithm week held in July 2004 in Humlebæk/Rungsted, two events were sponsored by FIRST: SWAT 2004 (the 9th Scandinavian Workshop on Algorithm Theory) and the Summer School on Experimental Algorithmics. In total, SWAT 2004 attracted 91 participants (including 2 invited speakers) and the summer school 20 (including 3 invited speakers). I want to use this opportunity to thank FIRST for the support we received. Based on the written and oral feedback, SWAT 2004 was an unmitigated success. In its entirety the scientific program was solid and of high quality. The invited talks by Charles E. Leiserson (on multithreaded algorithms using Cilk) and Gerth Stølting Brodal (on cache-oblivious algorithms) generated interest and discussion. In addition to the invited talks, 40 papers were presented. All the invited and refereed contributions can be found from the proceedings published in the Springer-Verlag series Lecture Notes in Computer Science (LNCS 3111). Also, the social program went well. The location (the Louisiana Museum of Modern Art) and the hotel (Rungstedgaard) were praised. When we decided to hold the traditional city-hall reception in Helsingør, we did not know the entertainer qualities of the mayor (Per Tærsbøl). As one of the participants put it "the meeting with the mayor was very very fun!". People will remember us from the relax atmosphere. In particular, the conference dinner was enjoyable. The concept of self-entertainment worked extraordinarily well, and it was a great pleasure to see all the creativity when people were solving our artistic exercises. The summer school was on my own research area so it is natural that I found the talks stimulating and interesting. Alexander Stepanov's message was that at the universities the senior faculty members doing experimental work should program more, and let the students write the papers, not vice versa as the tradition is today. He also called for a forum where one could publish programs, like in the Communications of the ACM in the sixties. Peter Sanders pointed out, among the other things, that branch misspredictions are significant for the performance of programs run on contemporary computers. For example, these determine the efficiency of most integer sorting programs. Hervé Brönnimann proposed a design for a geometric algorithm library which, after some incubation period, may be useful for other teachers as well. In addition to the aforementioned invited speakers, most of the other participants gave a talk too. All the invited talks were recorded on videotape, and the slides shown by the speakers are available via our website http://www.diku.dk/~jyrki/Sommerskole/. As to the social program of the summer school, we started with a dinner where after we watched the European soccer-championship final on a big screen. My Greek colleagues, who sat beside me, were very happy when their team won. For the barbecue the weather could have worked out a bit better; we had to eat inside. When organizing the algorithm week we produced software that may be useful for other conference organizers as well. These include: 1) a submission server that can be used to receive submissions and to support the program-committee work; 2) a web-based conference shop that can accept credit-card payments; and 3) a script that generates, automatically from the author contributions, the output required by Springer-Verlag for producing LNCS proceedings. I thank once more all of the many persons who took part in the organization of the algorithm week and provided this service for the whole algorithm-research community. Jyrki Katajainen jyrki@diku.dk