Hi,
Markus Neifer wrote:
> Java 3D is a 'high level' API and VRML is a 3D data format.
> How could you compare Java 3D and VRML?
Because you have authoring tools to generate VRML easily (once you know
how to deal
whith a product like 3DS max :-) : you can build quickly geometries and
simple behaviors
which can be loaded in a java3D app as a basis for a more complex
program. So, if someone want to create
a rather small 3D application he may compare the time spent to build the
scene in VRML using
an authoring tool and the time spent to build a Java3D application.
By comparing Java3D and VRML you have the opportunity the emphasize the
Java3D functionalties you
don't have in VRML.
> You may compare a Java 3D Viewer like VRML97Player and a VRML Viewer
> like CosmoPlayer but i would avoid that.
Why not ? Both seems to offer the same (partial) functionnality. Having
a benchmark with
complex VRML files could just show the performance enhancement of
Java3D. The problem could be that users
would prefer to use CosmoPlayer + EAI rather than using Java3D if sun
would not explicitely
show the benefit of using its product.
> Today the Java 3D stuff is still under construction with a lack of
> performance especially for large scenes.
As things can only go better for Java3D (otherwize we wouldn't be on
this list ;^)),
it is not a problem of giving a "bad advertising" of java3D. Remember
that
when Cosmo was in version 1, it wasn't a "killer" viewer. You can rather
show specific 3D stuff
which is better rendered with java3D (or better, which cannot be
renderer by VRML) to convince people of the JAVA3D advantages. I am sure
that a lot of Java3D functionalities can be demonstrated like that.
(And IMHO, the demos provided by sun are cool as code examples but
unfortunately not so accurate
to promote java3D).
cheers
renaud
