Dear friend Bertrand posts a rebuttal test to my Flow performance observations of yesterday, showing Flow performance isn't necessarily bad. Of course, his Flow example doesn't do much, and I haven't explained a lot about what ours does. Suffice to say that ours is a bit more complicated, but not in the JavaScript sense: all of the hard stuff actually happens in some Java components that are used by both the FlowScript and the Action version. If you put the two versions side-by-side, it's almost line-by-line identical code, apart from syntax of course. Still, we have a 60 ms penalty in the JavaScript version.