That's hardly conclusive proof... There's no indication of how they measured the time, not to mention if you load a program once, then close it and load it a second time, it will tend to load much quicker the second time.
Furthermore, app startup time isn't really an important statistic to measure. Unless you're closing and opening the app several times a day, in which case I'd question why you're closing it at all. Otherwise, it's a one time deal, so what is 10 seconds over the course of a day? The important thing to measure, is performance once it's loaded. In which case, assuming you have ample RAM, the program will rarely be reading from disk. If you don't have ample RAM, and you're using a lot of swap space, then that's a whole other issue to address. Defragmenting might help, but you'll get significantly better results by adding more RAM.
There's just too many unknowns from what I see in that discussion for any worthwhile conclusions to be drawn. I also think I mentioned high end audio and video editing as being one of the exceptions to the rule.
Simple fact is, defragmenting will only improve performance of disk intensive apps. Most programs your average user will be using on a day to day basis, do NOT fit into this category. Web browsers, word processors, email clients, MP3 players... None of these are disk intensive. Games can be, but usually only in the initial level loading, which is fleeting. And games are recreational, and thus could be said are a waste of time to begin with, so what's the point of trying to save time you're going to waste anyway? If you live in South Korea and are a professional StarCraft player, then maybe my comments don't hold the same value, but raise your hand if you are a professional gamer.