A Technical History of Apple's Operating Systems

I've been juggling numbers and I can't see where you got 12.5%. 1MB is 16.666...% less than 1.2MB, regardless of whether your MB is metric or binary, as long as it's the same MB for both figures. If the 10Mb is metric and the 1MB is binary (which is how I'd measure it), 1MB is 16.11392% less than 10Mb. Using binary 10Mb and metric 1MB gives 23.706+%. I've been unable to find a combination that gets below 16%. Those are measuring the difference as a percentage of the larger value (which seems appropriate for saying "A is X% less than B"). Measuring as percentage of the smaller value gives even larger percentages (20%, 19.2+%, 31+%, respectively). Could you elucidate? But yes, what I am saying is that, in practice, at least in my experience, the various levels of overhead eat up approximately 16% of the raw bandwidth. This is not a carefully measured figure; the uncertainty is high enough that it's more scientifically accurate to say "between 10% and 20%". And, of course, this is best-case. /~\ The ASCII der Mouse \ / Ribbon Campaign X Against HTML mouse at rodents.montreal.qc.ca / \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B