One way to eliminate the sliver is to take the app full-screen (if it will go full screen). Another is to pull the window down so that the bottom edge is below the screen boundary.
This is intentional, and it's not likely to go away until they redesign the GUI (and probably not even then). It's there because this is a graphical user interface, and even subtle cues like a one- to three-pixel gap can communicate important info.
There's a need to show that a window's edge is fully within the screen boundaries - that you're seeing everything there is to see. In Mac OS they use the shadow effect and gaps to show that. For example, the boundary between the grey menu bar at the top of the screen, and the gray bar at the top of a window - that's a gap. They could have drawn a one-pixel black border around every window (they used to, before OS X), but is there any functional difference between a one-pixel gap beyond the edge of the object, and a one-pixel line within that border? Either way, a row/column is being used for no other purpose than to indicate the boundary.
The only remaining quibble is whether a three-pixel gap at the bottom edge is necessary, or whether one pixel would suffice. I think this is due to the mouse pointer. By design, the tip of the mouse pointer can never go beyond the edge of the screen (unless it's a multi-monitor setup and the pointer is moving to an adjoining screen). Pull the mouse pointer to the bottom of the screen. You'll see the remaining bit is about three pixels tall. If it was just one pixel tall, it would just seem like a stray black or white pixel. At three pixels, it's discernible - you can see the mouse pointer below the bottom edge of the window. You'll see the same on the right-hand border of the screen - there's about three pixels still visible. On the left and top edges, the pointer extends right to the screen border, as the tip must reach every coordinate on screen, yet it cannot go beyond the edge of the screen.
