[EMAIL PROTECTED] writes: > instead of interleaving '1 spam/1 ham', do it based on # of ham vs > spam, ie: '9 spam/1 ham', etc. this will give a better simulation > than 'mixed ham/spam for a short time, then all spam' ...
Excellent. :-) It's a step up, but doesn't quite interleave evenly (think: ratio of 1.49). I have a feeling there's a clever algorithm for this somewhere... Daniel -- Daniel Quinlan http://www.pathname.com/~quinlan/
