-=> Quoting hugo landsman to Tom Almy <=-

 TA> However it does have a JIT compiler that
 TA> boosts performance, so with the benchmark I was using, NT is now 3x
 TA> faster than OS/2 in Java execution (was 2x).

 hl> Might be a coincedence, but to me such figures sound like the
 hl> improvements you might expect from running JAVA insecurely...

I wouldn't know about that. The program doesn't do anything but write to the
display. I can't believe that a secure interpreter could be 3x slower.

I just believe that the JIT compiler in OS/2 is doing a very poor job.

Secure or not, I can't believe Java will "save" OS/2 if it's 3x slower than
on Windows NT 3.51, which has a reputation as a hog.

Meanwhile I'm sitting here with IBM's free but "broken" Java port, and *3*
different Java ports for NT (Sun's, Borland's JIT compiler, and Microsoft's
(boo! hiss!)), two of which are also free, as far as I can tell. I can't see
where OS/2 has any sort of advantage here at all.


 hl> (BTW,
 hl> don't start telling us about NT's C2 ratings if-and-only-if *it is not
 hl> connected to anything*: obviously, anything is A1 when shut down:).

My system is only as secure as the lock on the door. And I'm nervous about
browser "cookies". But that's another story altogether.

Tom


--- Blue Wave/DOS v2.30