Because of differences like this ... I believe that the work done on HipHop represents a fundamentally smaller challenge than that taken on by the teams working to improve the implementations of languages like Python, Ruby, or Javascript.
I think this is true. On the other hand, it's not Facebook's job to do research. If they can benefit equally from solving a simpler problem, then that's what they should do.
I agree with the author's points, I just want to make explicit that Facebook didn't make wrong choices. It's just that their work is probably of limited value to everyone else.
There are many more people working on PHP apps than there are people working on dynamic language runtimes. Facebook's work is of limited value to everyone else for a limited value of 'everyone else'.
I'm taking the author's arguments at face value, that their project is of limited value to others because most PHP apps aren't CPU bound, and that they only handle a subset of the language.
I think this is true. On the other hand, it's not Facebook's job to do research. If they can benefit equally from solving a simpler problem, then that's what they should do.
I agree with the author's points, I just want to make explicit that Facebook didn't make wrong choices. It's just that their work is probably of limited value to everyone else.