On Mar 6, 11:38=C2=A0am, mdj  wrote:
> On Mar 6, 3:46=C2=A0am, apple2fr...{at}gmail.com wrote:
>
> > On Mar 5, 11:46=C2=A0pm, sicklittlemonkey
 wro=
te:> On Mar 5, 8:35=C2=A0pm, apple2fr...{at}gmail.com wrote:
>
> > > > This is true in theory; however in practice I have yet
to experienc=
e
> > > > any Java program running more quickly (or with less
memory) than an
> > > > equivalent C++ program on any platform.
>
> > > Experience away. I hope you'll excuse a C example:http://bytonic.de/h=
tml/benchmarks.html
> > > (Top right data cell.)
>
> > How many benchmarks did you have to search through before you came
> > across one that showed Java being faster than C or C++? =C2=A0I'll bet
> > quite a few.
>
> > Tryhttp://shootout.alioth.debian.org/gp4/benchmark.php?test=3Dall=E2=8C=
=A9=3Djav...
> > for a better comparison.
>
> To the extent that these benchmarks demonstrate anything useful (which
> is debatable), they clearly show that the speed differences between
> the two aren't very significant. C++ tends to come out on top by a
> small margin in most, Java in some. As a selection criteria, it's
> useless.
>
If you bother to visit the site and compare, for example, the Java 6
client (with JIT compiler) to C++ Intel, you will note the following:

For all 17 benchmarks, Java is never faster than C++, and up to 10x
slower.

For all 17 benchmarks, Java uses from 4 to 35 times more memory than
the corresponding C++ program.  In Java's defense, these are small
benchmarks, as the percentage difference would become smaller for
bigger programs.

> As for memory usage, Java is always worse, sometimes by an order of
> magnitude and then some. Fortunately though, this difference is only
> extreme in very small programs (which don't use much memory anyway)
> and the gap closes to a few percentage points as the application size
> grows. As a selection criteria, it's also useless.
>
Here I believe you are wrong.  I just loaded up Pages 2008 (Apple's
version of MS Word) on my (Intel) Mac.  It has a RSS of 77 megabytes,
and is, I believe, implemented in Objective C (which is more-or-less
equivalent to C++).  Next, I loaded up OpenOffice 3.0 (which is
implemented in Java) and selected the word processor.  It has a RSS of
130 megabytes, which is 69% larger.  Both pieces of software implement
approximately equivalent sets of functionality.  This is far from the
few percent as you suggest.

> That about runs us out of commercial reasons to prefer C++ over Java.
> Of course, for personal projects there's always aesthetics, and it's
> pretty obvious we have some different ideas there ;-)
>
Language selection will always be a religious issue.  C++ has
advantages in some areas, and Java in others (especially from a
management point of view).

I prefer languages that were designed by programmers for programmers,
and are not infected with ideas from committees or management.
Ultimately, however, any halfway decent programmer should be able to
program in any computer language given a few days to come up to speed
with it.  Sadly, such a yardstick probably rules out more than 90% of
people who call themselves programmers today.

--
Apple2Freak
--- SBBSecho 2.12-Win32