Occasionally I see people try to "optimize" their update statements to avoid writing the same value to a particular column. My understanding has always been that if you're going to update a row, assuming all of the values are in-row, the costs of locking the row are much higher than the incremental cost of updating one, two or all columns in that row.