Re: How hard would be an OpenSource JVM ?

Wilhelm Fitzpatrick writes:

> Secondly, creating a non-naive compiler/interpreter that executes Java 
> binaries with acceptable performance is not a trivial task either.  
> That's what all the projects you cited above have been working toward 
> for some time now.  As you observed, they have had some success, but 
> they are not all the way there.

Thirdly, Java is a moving target.  Suppose you started from scratch today; 
took the existing Java specs; threw a large development team at it, and good 
funding.

Say, for argument's sake, that two years from now you've completed and 
debugged a working VM, compiler, and a complete set of class libraries.

You know what?  You'll be back not exactly to square one, but pretty close.  
Why?  Because in the two years the Java class libraries would be twice as 
big; and the language would likely have sprouted more features.

Lather.  Rinse.  Repeat.

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQBBNlN4x9p3GYHlUOIRApciAJ4pOf2se5H73znWnP81UijixAkmtgCfXUhf
5S5FcL0engXPG4koYKFzHwQ=
=LJc4
-----END PGP SIGNATURE-----