My main gripe with java-- as you mention-- is speed. In
order to create good-looking GUIs you use Swing, which is
probably the slowest GUI toolkit in widespread use. What
Java needs is a bytecode-to-machinecode converter.
Since all the JVM does is translate bytecode to native
machine code on the fly, throwing out the results after it's
done, why couldn't there be an option to save the
translation, so it could run faster in the future?
Since everyone running Java programs has a JVM anyway, if
the JVM came with a translator, one could simply distribute
the bytecode version and the end user could run the
translator and create a binary optimized for their machine.
Re: Mediocre level engine, wich is a good start for most..
Mediocre, yes, but you must realize that this is only in the 0.7(as of last saturday) stadium, which is hardly production-level quality. I chose Irrlicht a) for its ease of use and b) for its licensing: it's under a zlib-style license. I doubt you'll find a less restrictive license on any 3D engine. It ain't Torque, but it has a good features-to-ease-of-use ratio.
> I'v tried this engine myself to see what
> the fuzz is all about :-)<br>
> The engine (and techdemo provided) looks
> (at first) nice. It has some nice things
> in store to fiddle with. (wich is nice
> for beginners too)<br>
> However, after playing with it for some
> time, the engine began to show more and
> more negative setbacks... for
> instance... <br>the graphic render
> module is not that good, it lacks the
> quality AAA-class engines have. (it
> actually made me think that i was seeing
> a render of a game back in
> So the engine is more an educational
> kind of engine, and not an engine usable
> for a commercial class game. <br>
> So if you want to be in 3d gaming
> business, this engine can proof itself,
> in the way of learning you how some
> things are accomplished in the 3d
> I'll stick with a more AAA class