Agreed re: pragmas as a hack, but that's just the kind of thing you'd expect to get to gain traction, before folding into a language standard. And re: data par

Agreed re: pragmas as a hack, but that's just the kind of thing you'd expect to get to gain traction, before folding into a language standard. And re: data parallelism, I don't see that anyone has made a terribly popular implementations. The niche languages like UPC, CAF, and HPF all seem to have withered on the vine. So far, the only thing people seem to buy into is that openmp-based task parallelism is easier than managing threads by yourself.