in reply to Recursive search for duplicate files
If you care about performance, you might just want to hash the first 5% (or the first 1k or whatever) and see if there are any collisions, and if there are you can still look at the entire file.
Frying pan on the stove
Results (327 votes). Check out past polls.