My personal experience is that duplicate records in the database means poor validation on both client and DB side as well as poor database design. I agree with those people who have gone the way of 'nip the evil in the bud' and not allow duplicates to get in the database.

The author mentioned that this article is for novice SQL Server users: I would rather train novice users how to avoid problems in the first place rather than react to them and realize that some things were left wanting in the DB design

Just my 2 cents