I should've posted this during OSCON, but you'll notice a distinct lack of posting in that period. :)
Anyway, the Python vs Parrot challenge is official, and the rules are simple.
The stakes are $10, a round of drinks, and a pie at 10 paces. Guido's not chosen his pie flavor yet. :)
The reasons for a few of those might be non-obvious, so some explanations.
#1 is because I want a benchmark, and don't know enough python to write one. Besides, this way there's no squawking about a bogus benchmark.
#2 is useful for both of us--me because I'll need to massage, Guido because he may have The Next Best Thing by then.
#3 so we get the latest Fast Thing in case either of us has done better since the last big release
#4 because this is a test of the interpreter speed, not our C compilers
#5 was Guido's but that's fine, I don't want to get penalized for disk latency
#6 is to cover the possibility that something fires off in the background -- it's not either of our fault if a cron daemon decides to go checking its data and blowing caches
#7 ought to be pretty obvious :)