From: MikkelFJ Date: 2001-12-04T04:21:36+09:00 Subject: [ruby-talk:27365] Featherweight concurrency and backends was: OCaml native code compiler has little to do with C-- > A few years later, Simon Peyton-Jones and Norman Ramsey had similar > ideas, but for a more source-language independent IL. We discussed it > for a while, and they liked the "C--" name, so I was happy to let them > use it. Their design is certainly more thorough and more modular than > what's in the ocamlopt compiler. Meanwhile I've done some more research on backends. There is Stackless Python which avoids the traditional stack callframes by using continuations. Continuations are also a key feature in C--. Personally, I think hardcore ruby should be based on C-- and steal from Stateless Python. Wrt. the .Net support - who knows maybe C-- will target MS-IL one day. That is of course only half the job of integrating with .NET, but anyway, I think C-- looks promising although immature. It may be a good idea with a bytecode compiler that can also generate C--. This is essentially the approach taken by OCaml - although that is a different C--. Create Ruby threads with continuations and kick ass :-) MikkelFJ Here are some of my personal Wiki links: Stackless Python Continuations background (pdf): http://www.stackless.com/spcpaper.pdf Stackless Python http://stackless.com/research/stackless/ C-- http://www.cminusminus.org Featherweight concurrency in portably assembly language Simon P. Jones on continuations in C-- http://www.cminusminus.org/abstracts/c--con.html http://www.cminusminus.org/download/c--con.pdf