On Sun, 10 Jan 1999 13:59:25 +0100, Spider <spi### [at] bahnhofse> wrote:

>
>
>"Ronald L. Parker" wrote:
>> You
>> could write an entire virtual machine and make the 'libraries' contain
>> p-code, but there would be a bit of a performance hit over just
>> compiling the functions into the source directly.
>Did I hear a JAVA whisper here, and a sob about the loss of Speed?

Only if you naively think VM==JAVA.  There have been bytecode compiled
machines around for decades. One such was UCSD Pascal.  Perl, too,
worked like this until the most recent version.  Java doesn't really
offer anything new in the arena of cross-platform operation, except
for the standardized interfaces in places like the AWT, and even that
is debatable with things like Tcl/Tk (and to a lesser extent Perl/Tk)

Be that as it may, it's actually not a bad idea, and I've been tossing
it around in the back of my head for a while now.  For the isosurface
stuff, at least, most of the work has already been done.  It's just a
matter of extending the existing formula mechanisms to handle looping
and conditionals, but I know from experience that extending those
mechanisms is not for the fainthearted.

The same principle applies to using Renderman-compliant shaders in
POV-Ray.  Again, you'd use a bytecode method (perhaps with a JIT
compiler on some platforms).  But again, you'd hardly use the Java VM.
It's designed to run Java, and it runs applications written in other
languages with less alacrity due to its choice of atomic operations.

Post a reply to this message