[zeRusski]

That was my suspicion as well: there could be a very usable and rich subset of Racket that could target JS without significant performance degradation. Sidestepping the deeper darker corners like I think ClojureScript does with Clojure? Herculean effort already it becomes even more daunting when you start picking up features to drop only to discover just how much of Racket depends on say delimeted control machinery. After all Racketeers have been eating their dog food so most if not all their PL research found its way into the design. Bitten by their own success :) The manifesto takes admirably healthy stance here: problems offer opportunity for research. Alas, when u consider the time it takes to build smth like Pyret or Whalesong your pool of researchers and implementers narrows pretty much to academia. Unless of course there's some benefactor who's willing to assume the expense and allow the time