* It's spatially an easier mental model.
* This mental model is consistent with iOS multitasking and iOS home screen.
* It allows for simple gestures to switch between desktops.
* In western society, we read from left to right. New is to the right. Fullscreen a new desktops are added to the right.
* This also sets up a linear "Forward/back" model similar to web pages.
* You can easily add ONE desktop on the fly. No need to make it a grid for no reason.
* The biggest reason: Spaces and expose are unified into Mission Control. One action to see EVERYTHING. This is a huge plus, as people don't have to check two different locations for a window. A grid of desktops wouldn't fit up above in mission control.
* There's nothing below the dock and there's nothing above the menubar. Simple mental model.
Regarding "maximize for windows":
* It's a "fit window to content" button. A window shouldn't take up more space than is necessary in a multi-tasking environment. Regardless, there is a fullscreen mode and windows also remember their position and dimensions if you drag them.