> My initial approach would be to use the 'cksum' command to create a > checksum string, and stick the pathnames into a hash indexed by > checksum. When you find a file with a checksum you've already seen, > you've got a duplicate. There's probably a Perl module that will do > a checksum so it can all be done within Perl. md5 is another tool like cksum, but the checksums it creates are more unique than that of cksum. David's method will work for everything, not just image files. What's cool about GraphicConverter, though, is not only that there's no programming on the user side required. GC can also find close duplicates. So if you have the same picture in different resolutions, GC can find it. If you have applied a softener to one picture, you can still find it. Oh, and two more things about GC: a) I am in no way affiliated with Lemkesoft, just a happy user. b) This feature happens to have been added on my request, just to give an example of how responsive the author is. Björn