Gregory Galperin (grg@ai.mit.edu) 2004-09-17 13:40:30.00 ISC Meeting 17SEP04 Cesar Cepeda = '06 Mexico (Chair SC) Gordon Cormack = elected '04; Canada Greg Galperin = '03 USA; outgoing (Chair '03 SC in USA) Ismail Hakki Toroslu = elected '03 (chair '99 SC in Turkey) Xenofon Papadopoulos = '04 Greece (Environment '04 in Greece) Krzysztof Stencel = '05 Poland: in charge of tasks Jyrki Nummenmaa = chair ITWG not present: Michalis Hatzopoulos = '04 Greece (Chair '04 SC) Tom Verhoeff = elected '02; Netherlands Requested that new members send an email to Tom including their PGP keys. Re-elected Tom as ISC Chair IOI05: Poland April 20-24, 2005 = planned ISC meeting * includes a Sat night for cheaper air fares * they intend to distribute tasks to the ISC in January - some ISC members want to stay untainted until April or so; they'll simply not look at the tasks until they're in the clear. - most won't work until the month before the ISC meeting * concurrent with IC meeting end of 2004 = external task submission => should have call for tasks by early November at the latest Note: they already have 10 tasks with test data prepared. will use their own grading system * they used 3 grading machines for 1000 person competition, think the 20 machines we had this year is overkill * this year, 146 submissions in last 2 minutes (need to compile & run) Krzysztof = environment & scoring; use their own system * linux & XP both, almost prepared * rhide (Gordon suggested devStudio, but too $$) no Java (Mexico not planning to have Java either, but open to discussion) IC meeting report by Jyrki: * Jyrki recommended to the IC that ISC should interview ISC candidates and even host candidates in the future. - We decided that the ISC should *not* endorse a candidate or speak against them - We should communicate our findings to the IC. (this would be especially valuable for choosing a host) - Therefore we have to somehow get the word out that countries bidding to host the IOI should bring someone to represent their Scientific Committee. Comments on IOI04: * 50% rule was successful, particularly on 2nd day * We might consider giving yet more information about test cases (e.g. list sizes of all input cases) * The ISC is pleased with the broad distribution of scores; medians are up substantially from previous years, which is very good. * Test data design is important; in the future we should pay more attention to this aspect of the competition. * Facilities for competition and for ISC were good, thanks to the Host. * Not all of the ISC Recommendations for hosting an IOI were followed, but the competition went well enough anyway. - Some of Host solutions were conceived well but not implemented as well as they could be, causing the limits and test data to be not entirely appropriate. - Test data were not verified, and there were several errors as a result. * There continue to be security issues at every IOI, including breaches with quarantine, computer security breaches, and issues with cheating. - Perhaps there should be a written recommendation on how to handle and enforce security? - Might work better if lodging for students is separate from leader lodging. - Security is another item to pay more attention to in the future. Jyrki volunteered to process the survey forms and have the data entered.