Re-2: [ObjectWeb architecture] just an idea for (Java) native compilation

[<markus.karg@xxxxxxxxx>]

> - Java still needs a native compiler.

Is this true actually? Have you done some latest benchmarking, comparing 
static native compilation (GCJ or similar) to the Tiger (1.5) or leatest 
Jikes, running on a Tiger (1.5) VM or the latest IBM VM?

I mean, what Java actually does right now IS native, optimized compilation, 
even if the compiling and optimizing is deferred to the VM. Actually this is 
what it makes possible to provide Tiger's Compiler API (let code be compiled 
and optimized at run time). The VM IS doing very good native compilation (e. 
g. it uses latest AMD and Intel instruction set extensions!).

So until there is a real world benchmark that showcases a valueable benefit 
from static precompilation compared to the LATEST (!) hotspot technology (not 
hotspot 1.3 or 1.4.2, but 1.5) I do not see that Java really needs a native 
compiler. 

And even if this would be the case: Java is controlled by JCP.org not by Sun. 
Why not just adding another JSR then instead of providing proprietary 
technology? I mean, if you don't like the government, do you leave the state? 
No, if you have trouble with Java, improve it by joining JCP.org! :-)

> So, as said before, at first thoughts, one could imagine the way 
> Java->OCaml->native code for compiling to Java to native code. If we think 
> only in a OO perspective, one could imagine a one-to-one mapping on Java 
> and 
> OCaml features.

Seems OCaml beeing reinvention of Java...?!

So if OCaml is so close to Java, maybe it would be better to write a 
OCaml-to-Java crosscompiler, then improving the Java compiler of upcomping 
1.6 or 1.7 release of Java by your knowledge on compiler technology?

Have Fun
Markus