[Haskell-cafe] Haskell on JVM

Simon, et al, It might be interesting to look at CALhttp://labs.businessobjects.com/cal/as a non-blank-slate beginning for Haskell on the JVM. To my mind there are three things that this needs to make it a real winner: - Much, much better Java interop. Basically, the bar to meet here is Scala/Java interop. - Better support for "std" Haskell syntax - and better support for some of the basic (semantic and syntactic) extensions - Figuring out what of Hackage it is reasonable to support Best wishes, --greg Date: Tue, 23 Jun 2009 15:16:03 +0100

From: Simon Peyton-Jones Subject: RE: [Haskell-cafe] Haskell on the iPhone To: Rick R , Haskell Cafe Message-ID: < 638ABD0A29C8884A91BC5FB5C349B1C33F4BAAFA41@EA-EXMSG-C334.europe.corp.microsoft.com

...

Content-Type: text/plain; charset="us-ascii"

Good news about the iPhone port!

There seems to be quite a bit more interest now in supporting platforms other than win/*nix on x86 these days*. Maybe now there will be sufficient motivation to make the fundamental changes required. Caveat: I have absolutely no idea of the scope or complexity of said changes. I will look through the LambdaVM code (and iPwn when available) to get an idea.

There is definitely an opportunity here for someone to make an impact. There is no reason in principle why Haskell can't run on a JVM or .net or other platform. But it's not just a simple matter of absorbing some patch or other. Here's a summary I wrote a little while ago:

http://haskell.org/haskellwiki/GHC:FAQ#Why_isn.27t_GHC_available_for_.NET_or...

The short summary is:

* There is interesting design work to do; and then interesting engineering work to make it a reality.

* We (at GHC HQ) would love to see that happen, but are not likely to drive it.

* If someone, or a small group, wanted to take up the cudgels and work on it, we'd be happy to advise.

* We'd certainly consider folding the results into the HEAD, provided the author(s) are willing to maintain it, and we feel that the result is comprehensible and maintainable.

* But another viable route might well be to use the GHC API, which means that the result wouldn't be part of GHC at all, just a client of the API. That would make it much easier to distribute upgrades etc, just as a Cabal package.

Simon

-- L.G. Meredith Managing Partner Biosimilarity LLC 1219 NW 83rd St Seattle, WA 98117 +1 206.650.3740 http://biosimilarity.blogspot.com