don said:

Al, Roger, Bird, anyone else, For your list of scannos, do you examine every instance of every word on the list? Many of those words would be more frequently correct than a misscan.

i recommend you not even consider testing for a specific scanno unless you suspect that it will return more hits than false-alarms. then i would collect data from every test, to _ensure_ that it does, and stop using any that do not. that's a very high bar to clear, but if a test returns more false-alarms than hits, it's wasting my time. it is more important and valuable to build an infrastructure which acts quickly and efficiently on error-reports from your community than to waste the time and energy of your digitization volunteers. sadly, p.g. never gave itself the opportunity to learn that lesson... -bowerbird