Simon Hengel <[EMAIL PROTECTED]> wrote: > > Definitely, characters. A high-granularity measure is essential to > > reduce the chance of ties. Even so there may well be equal-first-place > > winners -- hope they're not solved in terms of first submission, since > > submitting at 14:00 UTC is WAY easier for Europe residents (residents of > > the Americas would have to go to bed VERY late, get up VERY early, or > > spend extra effort setting up cron jobs), and that would bias everything > > in a most unfair manner. > > Not sure what to do about it, is there something more fair > than first come first serve?
Random choice between entries with identical number of characters would be more fair, because "first come first served" strongly favours European residents, given the timing at which entries are first accepted, as I mentioned. Time of entry could be made a factor if the geographical bias was removed (all entries in the first N hours of the context could be considered as "postmarked" at the same instant, with N computed so as to ensure some reasonable time in the evening for residents of Asia and the Americas -- for later entries, using time of submission is indeed fine). If you just sent PDFs of a nicely designed color certificate of "First Prize Ex Aequo" to all entrants who are tied for victory on the basis of number of characters, I would not mind if the keyboard went to the "first tied winner to submit" (since I'm not particularly keen to get the keyboard, just the fun and kudos of claiming I'm the/a winner, in the unlikely event my submission should be among the shortest;-). Alex -- http://mail.python.org/mailman/listinfo/python-list
