> Testing proves only the presence of errors not their absence.

I've never thought of it that way, but that totally makes sense.