Godmar Back writes:
> > Magician is a Java OpenGL API including implementations
> > for Linux, Win32, Mac and other OS, supporting JNI,
> > RNI, and Netscape's JRI. In contrast to Java3D, this
> > is a portable, popular, efficient solution.
> >From this perspective, I think that a Java binding for OpenGL, a JavaGL
> is a very useful thing to have, both as an agreed upon, open interface as
> well as an implementation. However, I do not believe it can or should
> supplant Java3D.
I said: Java3D is not a replacement for Java OpenGL bindings.
OpenGL is an immediate mode API, and Java3D is, in my
understanding, a scene graph API. There is no necessity,
and no advantage, in putting them into one proprietary
chunk. Layering Java3D on top of decent Java GL bindings
might be an option, having to use Java3D to get access
to hardware acceleration is not.
OpenGL is an open standard that can be cloned. The Free
Software community has Mesa (due to many years of dedicated
work by Brian Paul, and quite a bit of support from SGI).
XFree86 and others are putting together the bits and pieces
(GLX, Mesa, X11) to make hardware accelerated graphics a
reality for Linux, to name one important example. Java GL
bindings open the door for writing software that runs
on any platform that has OpenGL and Java - which removes
the barriers between Win32 and Linux/BSD/UNIX.
If Java3D becomes as widely available, fine. If somebody
writes an LGPL implementation of Java3D, more the power
to them. For many applications, a different API will still
be needed. I for one have no use for, and no interest in,
Java3D. I'd rather spend my days hacking my own JNI OpenGL
bindings, implementing the Magician API - which can also
be legally cloned.
b.
