Re: JVM Usenix paper got rejected

["Jonathan P . Olson" <olson at mmsi dot com>]

Sorry, didn't mean to offend anyone.  It's just that I've heard
a bit too much "Pure Java" hype lately.  Seems like any time I need
to get something difficult done I always need to write native code
and JNI isn't exactly friendly or type safe.

Personally, I've done alot of work with CNI and GCJ on embedded devices
and far prefer it when there's alot of C++/Java integration to do.

I recognize that it does constrain the Java implementation and garbage
collection algorithms, and is totally unsuited for some of the more
conventional
Java uses like applets, browsers, and the like.

But it also has a big win in compile-time type safety, speed, and efficiency.
And it sure is neat to subclass a Java object from C++ and actually have it
work!

On Fri, 15 Dec 2000 14:58:08 Nic Ferrier wrote:
> >>> "Jonathan P . Olson" <olson@mmsi.com> 15-Dec-00 6:58:05 PM >>>
> 
> >Seems that the open-source community is interested 
> >in source compatibility and clean integration with C++, 
> >whereas the Sun/Java community is totally obsessed
> >with "Pure Java".  Unfortunately, the Java people still 
> >believe the myth that a virtual machine implementation 
> >of Java can be efficient for embedded or operating systems 
> >work.  
> 
> Couldn't let this one go by... (sorry)
> 
> I'm one of those "Sun Java" people. And I don't think that.
> 
> I think that GCJ is a really usefull weapon in the armoury. It may
> even become a standard way of improving Java performance but my own
> view is it's likely to remain just another option.
> 
> A *lot* of work is happening on this and there are some quite
> interesting developments. GCJ is one. The potential of HotSpot engines
> is another and approaches like Jazelle is yet another.
> 
> Also, I don't think that the usenix people necessarily represent the
> "java community".
> 
> 
> Nic Ferrier
>