Some intelligent comparison algorithm should be developed to tag them as possible duplicates of stories that have been published in the last N minutes/hours/da

Some intelligent comparison algorithm should be developed to tag them as possible duplicates of stories that have been published in the last N minutes/hours/days. It shouldn't be very tough to come up with a beta version of the algorithm. They're asked in most technical interviews and everyone answers them :D