dude aweeeessssooooommmee. i'm definitely going to check this out later. i had seen the YHC/JS backend a while back but it looked like progress there had seized up. this seems pretty new. any experience with it?
I haven't done too much with it yet, to be honest — things have been a bit crazy at work — but FRP seems like a really interesting potential approach to JS-based web interfaces, and I can understand the desire to write Haskell and then compile it to JS (the assembly of the web, as plenty of people in the last three years have said). FRP JavaScript doesn't seem to have taken off in a big way in the three years since Flapjax came out, but perhaps we'll see that change soon.
flapjax seemed cool, but I really just can't stand writing with JS syntax and all the baggage that comes with it.
my current JS project (actually objective-j with cappuccino) is weighing in at over 5000 semicolons, 80% of which is probably glue code. FRP (or bindings if it's imperative) would reduce a lot of it, but it wouldn't make JS any "safer" to write in. i'd still get goofy mistakes all over the place that you take for granted won't happen in haskell/ML/etc.
I looked at ocamljs which can use FROC but it looked like it had a pretty large function calling overhead. YHC javascript backend also looked like it had a large overhead.
I also checked out scheme2js (actually I wrote the first version of my JS app in it, but with no GUI, haha) but it relies very heavily on the rest of HOP to do useful stuff in a browser, which I would not be using. Also, debugging in it was a nightmare. Inria does mad cool stuff though.
