[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]


Quoth McLagan, Doug on Monday, 7 January:
: I'd like to know what people think about the views expressed in this column:
: http://www.pbs.org/cringely/pulpit/pulpit20011101.html

Pretty much bogus.  The followup to criticism was bogus too.
Consider, for example:

> C# will JIT'd in Windows. The dynamic optimizations are sure to be
> implemented. Being coupled to Windows is irrelevant. In fact, C#
> will have MORE optimization opportunity because it is tied to
> Windows and Intel. Heck, a C# runtime could get to the level of
> avoiding CPU pipeline stalls because it can be sure of what its
> running on. Java could never come close to that.

Umm.  But every Java JIT knows what platform it's running on, and
can optimize just as well as any other native code compiler.  The
man is fundamentally confused, if he can't even distinguish between
a portable byte-code, and an architecture-specific compiler.  In fact,
there's no particular reason why a JVM on Windows shouldn't just
free-ride on top of the same JIT compiler that is doing these
hypothetical wonderful things for C#.

Don't even get me started on the nonsense of comparing Java
performance to C, rather than more comparable C++, and then using that
to infer that C# ("well, the names are similar, after all")
performance will be dramatically superior to Java's.

It gets worse after that.  Bob is not a competent commentator.