My statement regarding refreshing development databases was meant to suggest that the data that is being developed against is current. Assuming no schema changes have occured, a restore is appropriate to refresh this data. It accomplishes two things, gives you up to date data and goes a step further in testing your disaster recovery strategies. I was not implying to recommend that you overwrite development schema changes by restoring production data. Perhaps the statement was worded wrong, but I was just implying that its always a good idea to be developing against a "real-world" data set that closely matches production. This article was meant to be a very simplistic and theoretical view at development environments and if it sparks this much discussion I consider it a success.