six degrees of submission ... er, compilation.

[Jeff Kesselman]

>Tool reuse:  Everything else being equal, there's obviously a significant
pull
>             these days to leveraging all the effort going into Java, say,
>             rather than re-inventing the virtual machine, bytecode compiler,
>             just-in-time compiler, GUI builder &tc &tc on your own -- and
>             then documenting it all :).  Java happens to architecturally
>             forbid some of the things I want to do, and be ill-adapted to
>             others, and came along after I got started, so ... *sigh*.

A note. I though this too, which was why I designed and wrotethe kelvin
suite, but Im now abandonign kelvin because with 1.1 soem evy very
imopoirtant (to me) holes in JAVA were filled.  The "reflection" library in
1.1 solved  most fo my architectrual issues.

Jsu ta nteot hat if all youve looedkat is 1.0 you might want to look at 1.1
again, its grown alot..

SUN btw is also promising a JIT (just in Time compielr) by the fall thaty
will make JAVA run nas fast as native c++... and Sun to date has missed
neither a promise nor a date o nANY of their JAVA stuff.


>
>| Im not 'definite' about going interpreted tho, although it seems
>| highly likely that im about to embark on my sixth complete rewrite :-/
>| 
>| Has anyone ever done an analysis comparing the above 6 methods? It would
>| be interesting to look at.
>
>I think I've implemented all of them at one time or another, and that
>the most one can say is that engineering involves lots of tradeoffs,
>and you have to consider things carefully one project at a time :).
>
> Cynbe
>
>