Uhm, I don't find anything unclear about the OPs request. It certainly doesn't provide any wiggle space to come up with a solution that makes it possible to delete data. Now, I don't BrowserUK realized the chance of collision is 1 if there are duplicated files when he suggested to hash on file content; but after morgon pointed this out, Anonymous Monk presented that as a feature. That was what I was objecting to, and then BrowserUK came back he couldn't see data loss is happening, using some assumptions on how the OPs environment may look like.

Instead if people had said "you know, morgon is right, hashing on content has a much higher chance of collisions than hashing on file names", instead of defending the bad advice, this long subthread wouldn't have happened.

But Perl is a religion, isn't? We don't admit mistakes or better points of view.