You said that by using components, you end up using a big mega-app for launching everything. While it is true that this can be done, it does not need to be done. Therefore, the only answer to "Doesn't this defeat the purpose of having a multitasking operating system?" is simply, no, it doesn't.
The purpose of component-based design is to split frontend interfaces from backend functionality. Features in a gnapster bonobo component that iain mentioned in his entry can only be worked out if an extensive component system is in place. Using components won't lose any functionality, it would just increase flexibility.
Why everything should become a component rather than a chuck of code in a library that the application using it calls directly.
Rather than forking gnapster, why don't you try to do what iain suggested to do with bonobolizing gnapster by just "using its calls directly"?
And when there are no pros, the answer is usually very simple to arive at: the flexibility of this particular thing has no purpose, and therefore sucks.
Wow, you need to get some clue, dude.