Is Parallel functional programming a blind alley?

John Shalf, the team leader of the Advanced Technologies Group at the US National Energy Research Scientific Computing Center, in this interview indicates that efforts would be refocused on functional programming as a promising avenue for parallel computation. This approach failed in the 80s and 90s, and I fear will continue to be a blind alley. My critique of parallel functional programming may be found on pages 42-47 in the main report (http://arxiv4.library.cornell.edu/abs/1005.5183)