The duplicate detector works well, but it would work even better if its string-matching were slightly less exact, so that printer-friendly URLs submitted after canonical URLs would be caught, and if more of the variants of Economist magazine URLs were caught as duplicates.