Minutes: International Scientific Committee Meeting 4 Time/Date: 10:15 -- 12:45, 11th May 2001 Place: University of Tampere, Tampere, Finland Status: Approved (July 20, 2001) Present: Isto Aho Host n Kyung-Yong Chwa Host n+1 Gyula Horvath Elected Erkki Makinen Guest (Host SC) Ian Munro Elected Jyrki Nummenmaa Host n Tom Verhoeff Elected (chair) Baolin Yin Host n-1 1. The meeting was opened by Tom Verhoeff. The entire ISC is present. 2. The agenda was approved. 3. The minutes were discussed and corrected (two names, 7.4, 8). Action on items 9 and 10 is still to be taken. 4. The general part of the review report was presented in draft form. Based on the subsequent discussion, a final version will be prepared later today. The competition material received earlier this week, was returned together with written comments. The task-specific review report will be prepared today and tomorrow. The ISC will continue to be involved in the preparation of IOI2001. In particular, Gyula and Tom are still involved as task authors. Gyula and Ian will proofread task descriptions in mid June. At least two members (Gyula, Tom) will arrive earlier at IOI2001 to assist in the final preparations. 5. It was felt that five days are needed and sufficient for the review. It was considered indispensible that the review took place on-site. If possible, all ISC members should be involved in the review, and three non-host members are the minimum. The review process worked out well. The only reservations concern the timing of the review: when to do it. Doing it later leaves less time for adjusting the preparation. Doing it earlier means that less material is ready for review. 6. Kyung-Yong explained that three groups of people are already busy in Korea preparing the competition of IOI2002. These groups focus on the tasks, grading, and systems, respectively. They come from various universities throughout the country and meet every month. Some initiatives to obtain competition computers have already been taken. Kyung-Yong expressed concerns about using the same operating systems and grading system as at IOI2001. But he would postpone major decisions until after IOI2001. * Concerning IOI2003 in the USA, we expect that Rob Kolstad intends to use the grading system that he now is adapting for use at IOI2001. * We briefly discussed the choice of programming language(s) available at the IOI competition. It was observed that the pressure to allow Java has increased. For instance, Pascal has virtually disappeared from (university) text books (especially in the USA), and has been replaced mostly by Java. Still, it was felt that Java is not appropriate for the IOI. Three reasons were cited: - Unlike for Pascal, C, and C++, there is no vendor-independent approved standard for Java. Sun Microsystems Inc. has withdrawn Java from the standardization process. - Java is too complicated for use in secondary education. The object-oriented paradigm makes it much less accessible than Pascal or C. Furthermore, text input-output is too obscure. - In the IOI, Pascal and C must both still be retained. Adding Java as a third language is big burden for the organizers. - The speed differences between Pascal, C/C++, and Java are still such that the setting of time limits would be controversial. * The choice of operating system as development environment for the competitors should be evaluated at the end of IOI2001. It would be preferred to have only one operating system. * Concerning an IOI grading system, we very much wanted to see requirements and design documentation. The system should preferably also be easy to install and use for national olympiads. * Externally authored tasks appeared again at IOI2000 in China (three tasks), having been abandoned after IOI'93. For IOI2001,external authors submitted about eight tasks, of which a number are still under consideration. The contribution process still needs to be streamlined. It is not sufficiently clear who is responsible for what, and when and how responsibility is transferred to the Host SC. It is not reasonable to expect that authors complete a task fully ready for use in a competition. There are too many details that depend critically on parameters not known to or under control of the author. A three step delivery is suggested. In the first step, the author submits a PROPOSAL for a task, which includes a description, at least one solution program, an analysis of the task, and an indication of how the task can be graded. Based on this information, a decision is made about possible use of the task, and the author is informed on desirable changes and about constraints imposed by the competition environment. In the second step, the author completes the DESIGN for the task, which includes a careful wording of the task, alternative solution programs, more detailed grading information, etc. At that point, if the task is found appropriate, a member of the Host SC is assigned to the task. In the third step, the task IMPLEMENTATION is finalized. All details must then be completed, including experiments with "bad" programs. * The work of the IOI competitors is usually graded by executing the final implemention of their algorithms on a limited set of test cases, rather than by inspecting the program text and its design. This has the risk of portraying the IOI as a "hackers" contest, rather than a scientifically-oriented event. It is admitted that this situation may be undesirable, but also that it would be inpracticable to apply a different grading process (even though the other scientific olympiads do so). Nevertheless, it is important to maintain high scientific standards in preparing and executing an IOI competition. Tasks, solutions, test cases, and the grading process should all be prepared in a scientifically sound manner, and this should be properly advertised. * No ideas were offered for new task types. * The role of the ISC should be evaluated together with the GA. * The "Guidelines for IOI Competitions" should be updated in the light of the discussions during this week's review meeting. Especially Kyung-Yong expressed an interest in more details. Tom offered to start on this. 7. Kyung-Yong stated that the review meeting would have been more effective if the implementation of the IOI2001 competition material had been closer to completion. Jyrki responded that that had been his intention, but that several events had interfered, such as a change of sponsor for the computers, the late delivery of the computers for a lab to set up a small system, and troubles installing the software and configuring the systems. The computers had only arrived mid April. 8. Tom closes the meeting.