On Nov 17, 8:34 am, Miroslav Pokorny <[email protected]>
wrote:
> Vms that host both for example the CLR and JVM is nonsense due to the way
> each handles classloading for starters. They also use different byte codes,
> which would mean one or both would need to be emulated/translated into the
> instruction set of the other. What you are proposing has never happened in
> the cpu business, even when the Mac left 68000 and started to use power the
> older system had to be emulated. There is no previous example when two
> different systems with different instructions set somehow coexisted as first
> class citizens on a new execution environment.

The CLR and JVM are more similar than they are different though. I
don't really think it makes sense to compare native instruction sets,
this is after all a designated virtual layer with quite a bit more
flexibility. Even if we do focus on hardware however, remember that an
x86 RISC processor design is really just emulating the CISC
instructions in a fast and intelligent pre-fetcher.

Also, the fact that Google can write a completely independent register
based VM via manipulating stack based JVM bytecode, goes to show that
at least one way mapping is possible.

> I dont think many will be happy to adopt a system such as you mention (think
> like IKVM in mono land ) where their application is emulated inside another
> emulated system. Im no expert but having a double layered system of
> emulation must be quite slow no matter how great jit/aot becomes.

Yeah well, remember when Java first came out? It wasn't exactly
setting any speed records, and took some 10 years with a lot of
assistance from smart people as well as Moore's law to be considered
fast.

> Why would anyone trust Parrot ? If you thought Java 7 was slow look how long
> it took for Parrot to get their first release out... Do you really want to
> trust the world of computing on something that is that slow to in terms of
> evolution and proven ability to deliver on time ?

I especially DON'T want to trust any single VM. This whole topic
brought forth by Carl Joki revolves around raising the bar of
abstraction, to not be locked into one particular language nor VM
profile. I'm not saying it's easy, I'm saying it feels like a logical
next step in a world that needs to interoperate. Would I love to
replace my current heavy-weight server JVM, launching applets in the
browser of my laptop? Absolutely... Dalvik sounds like a perfect fit
for that.

-- 
You received this message because you are subscribed to the Google Groups "The 
Java Posse" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/javaposse?hl=en.

Reply via email to