Dakiu - Thanks, man - I'm glad you're liking it! I've updated the main post so it's easier to see - thanks for mentioning that.
In response to your second point, zoom to contain dynamically zooms the camera in or out so that all followed objects remain on the screen. The margins are the amount of space between the edges of the screen and the corresponding edges of the nearest followed objects - it acts like a border around the objects, essentially. The second layout in the example file gives a good idea of what's going on with the idea.