On 17-May-06, at 4:09 PM, Sebastián Sastre wrote:


> 	I can think a reason that could do that: to marry Smalltalk with
> assembler. A Smalltalk that makes it's compilations to machine code.
That's (indirectly) been done since 1984 or thereabouts. Peter  
Deutsch and Allan Schiffman did the work at PARQ and wrote a very  
famous paper for POPL (L. Peter Deutsch and Allan M. Schiffman,  
"Efficient Implementation of the Smalltalk-80 System", Conference  
Record of the Eleventh Annual ACM Symposium on Principles of  
Programming Languages, January 1984, pp. 297--302)
This was a specialized VM targetted at the 68k architecture and a  
couple of years later it was redone and extended and made portable as  
part of developing ParcPlace's HPS VM. At various times that has run  
(to my knowledge) on 68k, 386/486/pentium/blahblah, PPC, POWER, ARM,  
MIPS under a wide range of OSs. I did the first working ARM version  
in '94.
HPS compiles Smalltalk to bytecodes and then the VM converts those to  
quite well optimised machine code at need; it was a JIT many years  
before the javanauts made the term popular.

Compiling straight to machine code is certainly doable; it simply  
involves a lot more work since you have to develop and optimise and  
debug a *lot* more stuff. For example, you'd have to rewrite the  
compiler, the debugger, the InstructionStream related classes and  
tools, any system that expects to write out methods, etc etc. Send  
enough money and I will arrange it for you. Discussions could start  
at, ooh, One *Million* Euros.

tim
--
tim Rowledge; tim at rowledge.org; http://www.rowledge.org/tim
Useful Latin Phrases:- Gramen artificiosum odi = I hate Astroturf.