Alan Kay's list of recommended reading
squeakland.org
Something I find fascinating about this list is how diverse it is. Out of 100 or so books, only five relate directly to computers. The others come from a multitude of fields. I suspect such breadth is common amongst most of the greatest software innovators - people like Engelbart, Kay, etc. Knowing how to program is useful, but without breadth (and some depth) in other areas, your innovations are unlikely to have depth of their own.