Xith3D Reputation on the web

Engines, Libraries and Tools Xith3D Forums
#1

<<Xith3D is the slowest, its API is a bit a mess, lots of the code is broken. It just looks like it is an unfinished attempt, stopped in the middle of a global refactoring.>> – Niwak, on vbousquet.free.fr
Hopefully this has been mostly fixed and is in the process of being even better…

<<August 5, 2006 - 12:32 PM
The initial beta release of the RPG Toolset has been posted on the Download page. This version is, of course, nowhere near a completely functional toolset, but it’s a step towards that goal. At this point, I am considering switching from Xith3D over to jMonkey Engine, as I believe it will make a lot of things easier to implement. If you have any questions, comments, or bugs to report, please send me a message using the Contact page.

August 19, 2006 - 8:23 PM
I decided to go ahead and switch from Xith3D over to jMonkey Engine, so I’ve been working on redesigning things for the toolset. I have also taken the opportunity to clean up a lot of the interface code that’s already been written. A lot will have to be redone in the conversion process, but I believe it will help out in the long run. Just for fun I implemented shadows, and they definitely add a nice touch to the maps. At this point I am working towards version 0.2, which will consist of the new engine running on jME along with a basic tile editor and object editor and some other minor features here and there.>> http://www.rpgtoolset.net/index.php?page=home

http://blogs.vislab.usyd.edu.au/index.php/FrancisNaoum?cat=53 At least some users…

We have a page on OpenGL.org ?? Cool ! http://www.opengl.org/products/detail/xith3d/

I did not know William was still actively working on it ;D “Will is one of the two current lead developers of Xith3D.” (http://www.williamdenniss.com/)

Apple plays the compatibility card : <<We all know that Java3D had a cost, so now lets show a benefit for having
Java3D? Even the people who used to use it all acknowledge that it is a dead
API outside of some very specialized use cases. I know for certain that I
have fielded a significant number of queries from people trying to move away
from it to JOGL on OSX (which led to some bug fixes on my part) or Xith3D.
JOGL and Xith3D are both faster, more capable, and more compatible. Xith3D
also offers API compatibility at the 95%.>> http://lists.apple.com/archives/java-dev/2004/Jul/msg00030.html

<<I’m planning to stick to higher-level API’s now, as low-level just has too many quirks that need to be figured out. Would you recommend learning Xith3D or jME? I know jME only support JLWGL, but Xith3D can support both JOGL and JLWGL, and plus JOGL integrates very cleanly into AWT and Swing. I don’t know if Xith can do all that jME does though, such as the 2d text rendering that digiwired was talking about and the buttons. If both of their functionality is more or less the same and the only cost is a bit of performance, I would rather go with Xith for the added compatibility with Swing.>>
<<I’m gonna have to learn either jMe or Xith>>
<<jME definetly :smiley:

Xith3D is based on Java3D (slow, horrible…etc) so Xith3D comes from that line

jME is based on NetImmerse(a professional development suite) which is already known and tested. Collision detection is extremely fast in jME. And im developing an AI System for jME too (not part of the scenegraph but a different jar); so you get your AI tightly integrated with the scenegraph too. I dont believe Xith3D has an AI System.>> (He’s probably one of those who believes a scenegraph should prepare your meal… Aso, what does he mean with an AI System?)

<> http://gcc.gnu.org/ml/java/2004-06/msg00417.html

<<Personally, I had explored Java3D before this. While it has its place, Java3D is really too slow for, and just isn’t geared for, game development. Because I wanted to work on games, this caused me to look elsewhere, which led me to the Xith3D project. I went from there to jME, where I have stayed.

What initially attracted me to jME (away from Xith3D) was the highly active development community. I also liked that jME wasn’t trying to mimic another API. Whether that’s fair or not, it gave me the feeling of additional freedom and alliviated some concerns.>>
<<As far as Java Scene Graph API’s go, jME is the way to go. The community is awesome and the API is very easy to use and straightforward. I believe it performs better than Xith3D and the documenation is better and improving all the time.>>
http://www.devmaster.net/engines/engine_details.php?id=78

Tell me, what’s the use of Linux support ? ;D http://homepage.mac.com/dsouthwick/FileSharing18.html

I noticed that jME users were really agressive on their forums, e.g. “seems like jME is kicking XIth3D rear”. Heck, what did we do to them ??

#2

I must mention that there are lots of savage tales spread about Xith3D (not talking about FPS here, agreed) and Java3D and jME. Most of the ones talking on forums just talk by hearsay… “I have heard that Xith3D is 2x faster than Java3D”, “I have read that jME is 4x faster than Java3D”…

Also the web isn’t as reactive as it says… recent changes in Xith3D (perfs boosts, e.g.) have not been diffused outside from Xith3D forums… what? of course, if it’s not good news for jMe, then… there’s no reason to quote that, heh? ;D

However, these quotes gives a shot of what users were deceived in (a long time ago) :

  • Support (hopefully this is fixed now, I think between Marvin, Yuri and me, we provide a pretty good support)
  • Documentation : filled partly with the “Xith In a Nutshell” book, the website is being regularly improved
  • Performance : display lists and shaders/sorting/whatever are what we (mostly Marvin and Yuri) have been working on. Also thanks to Yuri faster texture loading is now available.
  • Bugs : more and more being squished all day, some constraining us to revise the architecture a bit (sometimes getting far from Java3D).
  • Features : pure HUD is constantly being added widgets/features (e.g. drag-and-drop), Marvin’s easy coding classes are a success (hawkwind, rafa_es, me can witness), copy-to-texture and render-to-texture are on the way, maybe some more work on advanced shaders would be good (at least high-level shaders have been implemented by Florian aka Goliat). Terrain code is being revised. Support for precomputed animations has been added, + Cal3D loader. COLLADA loader not to forget (Croft’s great work, can be found on whoola.sourceforge.net). And many, many more to come. (E.g. I’m considering an eclipse plugin for a HUD GUI builder).
#3

The problem is, that most of the comments are from 2005. So they were written before the new features like the HUD, wrapper classes and XIN. Which are probably the most important things for the very end user of Xith. And I guess community support was some kind of dead, which is a whole lot better nowadays.

Seems like we need a press agent to pronounce the improvements to the world ;).

Marvin

#4

Btw, digiwired is me on the java.net forums, that was my old nickname before I was christened DarkProphet by a friend.

Heh, thats special, sparky is trying to insult people. Lets see, how many scenegraphs have you worked on? Infact, how many engines have you started and finished yourself ? So before mouthing off at people you dont know; see if it applies to you first…

Go have a look at the AI section on the JME forums, you might learn a few things…

DP

#5

@Amos: It’s seems you have quite the chip on your shoulder. There isn’t a huge “bash Xith” party going on over at the jme forums, nor is that a goal. The only performance goal we have is to continually make jME faster and more feature rich today than it was yesterday. If people are moving to jME or saying they like their experience with us, that’s their opinion and they are entitled to it. I think I have personally proven that jME devs have been known to help the community at large; you yourself said as much. So calm down and bring your focus back to why you are working on Xith in the first place… because you like it and that reason alone is good enough for you. ( Why else would most people be working on a game or 3d engine in Java? :wink: )

#6

he means… an artificial intelligence system. a not-so-rare feature.

I doubt the forum is like that. from the bits I’ve seen, they’re more focused on their own product than competitors. you’re just reading that 1% of the forum that consists of fanbois.

you’re very self-involved in the xith project, which means you’re going to be defensive, but don’t fool yourself - jME has alot more ammo than Xith at this point, and Xith would not fare well in a 1-on-1 comparison yet. if this wasn’t true, you wouldn’t be working on Xith around the clock.

don’t worry about what the web is thinking about, just focus on your project and get things done. search for Xith quotes later after there’s a finished product to be quoted.

#7

Even if I haven’t had a look at the jME’s AI system, I wonder, how it could look like. I guess there had to be one for strategy, racing, RPG, adventure, etc. And each one must be highly generic to be able to be used by not only one project. If it is this way, great compliment ;D. And in this case it should be an indepentent project, not tied to any scenegraph, such that it could be used by any project, nevermind which scenegraph is used.

Marvin

#8

translation: “we’re not able to put one in Xith”

edit: maybe that was a bit harsh. but your post just seemed like you were coming up with a quick excuse to not include such a feature in Xith.

#9

Were you referring to my post ???
No it is not an excuse. We’Re working on so many things, the AI thing certainly is a less important one at the moment. It is only a kind suggestion to you jME guys to think about writing the AI system as an indepentent project, since in principle it is indepentent from any scenegraph stuff. Is yours tied to jME’s scenegraph in any way? It would be so cool, if we could use the AI system with projects written in Xith3D. Why writing two AI systems?

Marvin

#10

ah, i guess I misinterpreted then :wink:

#11

Indeed. Someone?

Wow, please dont feel offended. The “meal preparing thing” was more of a joke than an insult, really. Sorry if you took it bad.

One, Xith3D.

I have started long ago something named Gamma. I trashed it, though now I consider that I learned a lot from trying that.

Yeah sure, but I just don’t see what you do mean by “mouthing off”. And I have absolute no pretention about the crap I’m able to achieve.

Sure, reading on the web is nearly always instructive.

Yeah, you must be right about the “chip on the shoulder” thing. And I must agree with the remaining.

Well, seconding Marvin, I wonder how it’s done. And I sure would learn lots of things if I studied it.

Well, I hate to denounce, but http://www.jmonkeyengine.com/jmeforum/index.php?topic=2597.msg19967#msg19967 is an example (not too agressive, this one). Others examples of that kind of things is irrisor’s posts on my old “Game APIs in Java” thread. But important jME developers (understand : mojo, renanse) don’t do that.

Well, yeah, and…?

…and ok.

+1. That’s what I tried to subtly suggest in my “prepare your meal” joke.
But if integrating AI with a scenegraph is a real plus, then OK.

However, I do still believe it’s stupid to have competition like that. Competition may be good for users but in the end we just write the same things several times. And of course as soon as “merger” word appear somewhere, even if it’s to talk about a common effort, well that’s really, really, really, really, really not welcomed.