That's a "good" traditional article on the subject... with absolutely no mal intent directed towards you or that author (neither of you have done the experiments I have and you're both operating on "Best Practice" recommendations) the problem is that most such articles are based on supposed "Best Practices" and the testing I've been doing for about the last 3 months prove that such traditional methods actually cause performance problems because people wait too long to defrag or have a 100% fill factor on out of sorted order indexes.  In the former, all the page splits that have been held off suddenly happen on the day before it's detected for defrag and in the later, page splits go absolutely crazy just as soon as the defrag is done.  That is, unless the table has gotten large enough for the "Reorg" portion of most code to kick in and then it doesn't do any better than not defragging at all.