@vaporeon_ how do you feel about targeting LLVM bytecode
@vaporeon_ yes i mean the IR
@aescling Conceptually, it makes sense, you're just adding another front-end to a compiler that already knows how to do C, C++, FORTRAN, and probably a bunch of other languages
The front-end/back-end split of LLVM is kind of neat
However, it may be difficult to get that to run on the computers that I want to try Scheme on (e.g. SPARC hardware running Solaris 7)
@aescling Bytecode? Do you mean LLVM Intermediate Representation? Basically you write a new front-end for LLVM and then let the LLVM back-ends handle the actual compiling to machine-specific object code?