I think it's quite telling that in all of that text there are actually no real examples as to how a program will actually look like.

Call me old if you like, but I've seen countless of similar "manifestos" covering everything from flight simulators to, yes, programming languages. Only on very rare occasions do they result in something concrete.

>"Write once, parallelize anywhere" -- With very little modification, the same Flow code should be able to target hadoop / Google's MapReduce, the JVM using Java Threads, C using pthreads, MPI or a GPU using CUDA. By understanding the big-Oh complexity of computations in Flow programs, The compiler will produce code that optimally partitions the workload in systems with nonuniform computing speeds and/or communication costs, e.g. CPU/GPU combinations.

That's a big promise.

When I read through the text, it didn't once think "this can't work". It seems plausible and if implemented, this could indeed be a big step forward.

I hope this, or a similar effort, will be actually be realized.