From The Axis of Eval: "Truth IS stranger than fiction": Chuck McManis explains the cancellation and subsequent fame of Java.
It's fascinating to me how often you hear about these overnight successes that took years to develop and were canceled, given up for dead, etc. many times.
I remember hearing similar stories about Self and Erlang at HOPL-3, the History of Programming Languages conference. Maybe you shouldn't bet on the success of a new language unless it's been cancelled or abandoned at least once?