IOI2000 Questionnaire for Delegations Summary of further comments written on 63 returned questionnaires [Remarks in square brackets are mine, Tom Verhoeff] ======= Regarding process of accepting tasks: key question should be "did we produce a better quality of tasks". Needs more discussion. Procedure for accepting the tasks is not good, because there was not enough time to discuss the objections. Objection to non-secret input (output-only) tasks: that way you cannot measure running time or memory usage. [Run time and memory are naturally limited by the duration of the competition and by the competition computer hardwared.] Append author's solution to handout. Add more details to handouts. Final versions of tasks arrived much too late. Surprised that ISC had not found serious shortcomings in task WALLS. Let khamis@frcu.eun.eg know how to subscribe to IOI mailing list. Even better to distribute handouts at question time. Non-secret input tasks are very bad idea. [Why?] Distribute handouts 20 minutes after distributing tasks. Task acception process was not sufficiently democratic. Expand info in handout. A preliminary discussion on each task is necessary, before voting. Limit discussion time to 30 minutes per task. Discussion helps delegations getting a better perspective of the task. IOI is wonderful. A complete model solution with comments after IOI would be pedagogically interesting. In case of non-secret input tasks, the input data must be more complex than they were now. Although the process of accepting the tasks was easier, doesn't mean it was better. The fact that the committee took all the comments and decided what to use and how, without a chance to comment again, made the tasks, especially those of Day 2, particulary confusing in the final version. Only task WALLS was not good for translation. Concerning acceptance of tasks: With no discussion all is easy. None of the tasks would have been appropriate for non-secret input. Don't like the idea of non-secret input tasks. The input data has to be secret. Don't submit silly problems like BLOCK. I am uncomfortable with the fact that the ISC knows the tasks beforehand. Even with the ISC the quality of tasks is not better than in IOI'99. The process of accepting the tasks without some discussion is not very good. The voting without counting is not good too. Subscribe arw@mail.cmb.ac.lk to IOI mailing list. We are opposing the existence of non-secret input data type tasks. It would be appropriate to have the IOI mailing list browsable from the website. The SC Chair (generally) should appear to be more flexible. Only in task WALLS, there is some ambiguity related to the way to move from one region to another. Concerning task acceptance: - Seems poor that, when a large number objected to BLOCK, there was no way to get rid of it. To my mind, 15 votes against a task is as high as we have ever had. - Seemed poor that there was no opportunity to reject tasks "as a set". - No feedback on rejected task comments: would any ambiguities that the SC disagreed with ever get resolved? - English in modified tasks was poor, with no opportunity to comment new issues (other than through translation), such as new technical "special cases". Concerning moment for handouts, time is needed to think about solutions before meeting the competitors. Concerning non-secret input tasks: too many of the IOI2000 task inputs can be solved by hand (which may be acceptable for some simple test cases). This does not easily apply to the reactive task. Concerning submitting tasks for future IOIs: It is hard to submit if this would lead to a restriction on involvement with candidate competitors (especially in countries with small organizations). The rules on this should be more clearly stated.