[ Quoting Denis Tonn to Scott Little ]
 DT> Does the win95 logo printed on CPU's and all kinds of other hardware
 DT> make them "Windows 95 CPUs" or "Windows 95 hard drives" or "Windows 95
 DT> systems"?
I don't care what YOU think, there are Java CPUs which can nativly understand
Java code. 
 DT> I repeat, it takes much more that just a CPU to make a Java Virtual 
 DT> Machine.
VM? Why use a VM on a CPU that understands Java all by its lonesome?
 DT> enviroment it runs in. JIT compilers will compile the Java bytecodes
 DT> into "native" machine language, but these still need the JVM to run 
Correct. However, they [JIT] are about three times as slow as when being
executed on a Java processor.
 DT> systems or an application that takes 2 man years to write and only 
 DT> runs on one CPU and under one operating environment?
I'd rather wait two years and get the one that was written by a decent
programmer, that takes full advantage of ALL available resources. I do not 
want
a cheap, generic program, that is three times as slow as it would be if run 
n
a native processor.
 DT> applications are written in high level languages such as "C", Pascal,
 DT> BASIC, Cobol, Fortran, Smalltalk, (ad nauseam) than native assembler
The problem lies in both the interpreter/JIT and the language itself. Since
Sun's language, Java, does not contain anything specific to any CPU (note: I
said SUN'S JAVA. Those extras I was referring to before are not Sun's
additions), it puts a LOT of extra pressure on the interpreter to figure out
how to convert generic intructions into the most efficient code for that CPU.
 DT> "efficient" than assembler). I *LIKE* assembler and have coded a fair
 DT> amount in it, yet I still have more code produced in a high level 
Same here. Problem is, the program is IN a form the computer we're on can
understand. With Java bytecode, the computer must continually convert it
every time it's executed. It would make much more sense to kill JITs and keep
interpreters to hypertext media, and distribute converters instead. I'd 
ather
leave my computer converting bytecode into native code for a few hours, than
wait an extra 30 seconds for the program to load.
 DT> But Java is so much more than just "bytecodes" (like "Pcode" is), you
 DT> have to be careful with the comparisons. 
It's a good idea thats been given the FUBAR treatment. See my comments on
converters above.
Regards,
 - Scott
[ admin@cyberia.asstdc.com.au | www.asstdc.com.au/~cyberia ]
--- FMailX 1.22
3:712/848)
---------------