Report

Congratulations to Lund University, the regional champions!

The 2005 NWERC at KTH, Stockholm, 11-13 November went quite smoothly. This was thanks to the support of ACM, IBM and locally from KTH and a number of volunteers.

There were a total of 46 teams from Denmark, Finland, Germany, Great Britain, The Netherlands, Norway, and Sweden. The top twelve teams came from twelve different universities, representing all the participating countries. This illustrates that this regional contest is highly competitive.

Steering Committee

Mikael Goldmann, NWERC Director
Fredrik Niemelä Head Judge and NWERC Co-director
Gunnar Kreitz, Technical Director
Linda Kann, Head of hosts and runners
Per Austrin, Vice Head Judge
Mattias de Zalenski, Vice Technical Director
Henrik Eriksson, Judge

Judges

Fredrik Niemelä, KTH, Head Judge
Per Austrin, KTH
Henrik Eriksson, KTH
Øyvind Grotmol, Medallia, formerly NTNU
Gunnar Kreitz, KTH
Jan Kuipers, U Utrecht
Jimmy Mårdell, Elucidon, formerly Umeå universitet
Stein Norheim, Tieto Enator, formerly Linköpings universitet
David Rydh, KTH and Chalmers
Tom Verhoeff, TU Eindhoven

Problem set

The problem set contained nine problems that were developed over a period of about six months. Every team solved at least one problem, and the winners solved six problems. Problem A was not solved by anyone during the contest.

Team Contest Reference

NWERC has traditionally allowed any information on paper as reference materials during the contest. This year, to reflect the team notebook rules at the World Finals, the NWERC rules were updated. Each team was allowed a 25 page "Team Contest Reference" (TCR) in addition to on-line documentation of Java and C/C++. Up to three copies of the TCR were left at the team computer after the practice session for inspection by the judges, and were then available to the team on the next day. Paper, pens and pencils were provided by the organizers. Contestants were not allowed to bring anything except mascots and food to the team computer on the contest day.

Coaches Meeting

During the coaches meeting we mainly discussed how we could further improve the contest next year, and if we should continue using the TCR as the only documentation available to the teams apart from online documentation of Java and C/C++.

It was generally agreed that the talks on Saturday, while very interesting, had been a few too many. Next year we will make sure that there are fewer talks and that there is some other activity of a more recreational nature.

Regarding the TCR, most coaches favored this rule change. A few coaches said that while the information about the rule change had been available for a long time, the NWERC05 main page should have stressed that there was a significant change as compared to previous years, otherwise experienced coaches might just assume that there would be no reason to read the rules page. One coach had not realized the change until a few weeks before the contest. There were a couple of suggestions for the future. One was to provide a sample TCR, perhaps by asking teams if they would be willing to publish their TCRs on the web. This may be difficult as keeping a good TCR secret gives you an edge in the contest. Another suggestion was to allow three books as an alternative to the TCR. However, there was strong opposition to this, in part because it would deviate from WF rules and in part because it can be tricky to define what a "book" is, and it was decided that the rules regarding reference materials should be the same in 2006 as they were this year (provided that the WF rules do not change significantly).

Contest system

The CSUS PC² system was used to judge the submissions. It was configured for multi-site, but only in order to have a backup server in case the contest server would break down.

A contest specific installation of Solaris was used for security reasons and to ensure that local teams could not access their home directories. Since each contest computer ran as an independent workstation there was no contention for the network, file servers or other shared resources. The team computers could only communicate with the PC² server and a dedicated web server that provided the scoreboard and the programming language documentation. This worked very well and made it unnecessary to list exactly which software would be allowed during the contest. Unwanted software such as maple, lex and yacc was simply not included in the installation.

NWERC Hockey League

The NWERC Hockey League (NHL) was inspired by the Java Challenge at several previous World Finals. The NHL consists of an Ice Hockey simulator and a framework for defining the players of a hockey team. The NHL was announced about a month before the contest and the software could be downloaded from the NWERC web site.

The play-offs took place during the contest dinner on Saturday. While only six universities participated by sending in teams, most participants followed the contest, which provoked both cheers and lots of laughs.

The top teams were quite strong, and after a close contest the team from Lund had won the NHL.

We believe the main reason many universities did not participate in the NHL was that they overestimated the amount of work required and perhaps underestimated the amount of fun involved. We may try to repeat it next year and see if there is a better turnout then.

Thanks

ACM organizes a wonderful contest, and provides a tremendous amount of help to the regional organizers. The RCD web is very useful and has answers to a lot of questions. And when I could not find the answers, ICPC Manager was only an email away.

Thanks go to the sponsors. Without the support from IBM, the primary sponsor of ACM ICPC, NWERC 2005 would not have been possible. IBM Sweden has also provided much support, and we particularly thank Johan Ekesiöö, Örjan Granqvist, and Barbro Lådö. We also thank Medallia for sponsoring the contest dinner, and particularly we thank Øyvind Grotmol and Erling Ellingsen.

KTH has been very helpful in organizing the contest, and we particularly thank Anders Flodström, president of KTH, Ingrid Melinder, dean of the School of Computer Science and Communication (KTH CSC), and the computer systems staff at KTH CSC.

The jury and the steering committee did a great job, and we also had a number of volunteer hosts and runners: Alexander Baltatzis, Max Bennedich, Marcus Dicander, Andreas Ehn, Kristofer Karlson, Mikael Lagerkvist, Jon Larsson, Helena Lundvall, Stefan Nilsson, Pontus Stenetorp, Martin Wendin, Ola Westin, and Henrik Ygge. Thanks to all of you for making it happen!

Finally, Roy Andersson and Roger Henriksson at Lund University, organizers of NWERC 2003/2004, have answered a lot of questions and made sure that we stayed on schedule and on target a few times. Thanks guys!

Mikael Goldmann, NWERC Director
Fredrik Niemelä, NWERC Co-director