Back of the Virtual Envelope (Or Is that Virtually Back of the Envelope?)
Assume that it takes a graduate student only an hour to work through
all the detector data from an event and see if it is interesting or not.
Assume one event in a billion is interesting, and we have ten billion events.
This is 10^10 student-hours. Since these are, after all, graduate students,
we shall introduce 20 hour days and 365 day years, giving 1.4 * 10^6
student-years. Alexander the Great conquered the whole of the Persian
Empire with perhaps 30,000 men; let us say 35,000 people forms an ``army'',
since this makes the division nicer. One army of graduate students would take
40 years to process the data from a single experiment.
Sometimes the processing is much faster; I could scan a hundred bubble
chamber photos in an hour, in the palmy days of my youth. This would reduce
the time needed for analysis to half a year - if you employed a plurality of
the world's physicists under a regime unknown outside forced-labor camps, such
as the city of Manchester in 1844.
By contrast, actual experiments where only one event in a billion passed
all the cuts, on which no more than 600 physicists worked in all capacities,
and then likely not over sixty hours a week, have analysed their data in less
than a year, i.e., they had no more than about 200,000 hours available, and
probably used much less than that.