A. Nonymous said...

Engines, Libraries and Tools Xith3D Forums
#1

“IMHO, Xith3D lacks two things : Users and a Direction”

My guess is that once we found the direction and go toward it (with good doc as we’re actually writing), users will come along.

So what’s the direction ?

My will is that Xith3D ought to be a :

  • Portable 3D scenegraph using OpenGL
    which is
  • Free software (BSD)
    and
  • High-speed (we may have some work on that, but not that muich since it was designed for efficiency by David Yazel)
    and
  • Which ease game development and provides a convenient graphic pipeline e.g. and some behavior classes, all kinds of utils (and we should have a consensus on that cause now everyone’s going his own way : Yazel, croft, Qudus, me, others)
#2

No need to be coy, twas me, but it is polite to ask people before quoting them from private messages (be it anonymous or not).

I think what you’ve got there is goals, direction generally tell you how you’re going to get there. I think your development plan was a very good start - but it’s extremely low level, the sort of thing that Qudos at the speed he works could have done in a matter of days.

A good thing I think would be to have a game that was being developed against Xith, stressing it and showing the world that Xith can be used for great games development. I tried to do this with Martian Madness when Xith originally came out of the womb (birth references everywhere for me these days).

If you’re developing a game with Xith at the moment, show it, and show what great things it can do with Xith. Show it breaking Xith, show Xith being fixed. If a game is too big then some impressive demos. JavaCoolDude’s stuff was great for showing off Xith as looking fantastic, unfortunately the code either went missing or is extremely difficult to read (MD3 loading for example).

Direction is generally given by users, not the other way round, especially for for Open Source projects. The software is created because someone needs it for something they’re doing (i.e. Magicosm) and then they’re nice enough to Open Source it. Features that get added cause they’d be a nice to have tend to get implemented and never tested in a real world scenario (see Shadows, or BSP loading). Features that get added because someone wants to use it in their game automagically get tested.

Currently I’m working on a project which I did consider using Java3D and Xith for. I’ve opted for JME. Why?

Java3D - Great to see Sun supporting it again. Not convinced that it’s keeping up with features or being updated as often as some of the others.

Xith3D - Nice to see active forums again (after a period of quiet). Not confident it’s being driven forward by any users and don’t have the time (insert baby soon excuse again) to be that user.

For reference I also said in that private message that I believe Xith is based on good core software from the original project and that I have no doubt it’s perfectly good for games development. I’m just not sure where it’s going and what’s going to cause it to get there.

Kev

#3

I think, when we have reached all the goals of the dev-plan, we can talk of a stable and very good 3d engine. Amos and me once talked about a small game (like pacman or so) which we could create to show what can be done with xith. It should be a small game to not let it be too hard to understand it. Demos are always fine, but a game would be better. I’ll add an additional point to the dev-plan.

#4

Well I have (I think ???) the game. I just need to put it where it can be played, preferably webstarted. I have been talking with BlueSky on this. All ideas/options appreciated.

So…first its is based on an early version of Xith. Only for API/collision stability. I am now upgrading to the current Xith release.

The game includes…

o multiple rooms, some hidden
o pick activating lighting
o pick activated doors, keys, notes, UI dialog, etc.
o UI is based on the Xith swing UI, I have popup dialogs, notes, etc
o UI based game dialogs, color match, puzzles combo locks etc.
o Animated textures
o Javacooldude particles
o Collisions including gravity based effects using old Xith collision system
o speech synthethise using FreeTTS. Yup the dang thing can talk…yeee haaa!! ;D

What I need is a way to present this to the community. I don’t know swat about jnlp files. I don’t have a personal site to deliver my work from :-[.

#5

Sounds excellent :slight_smile: Can we see some screenshots at least? That might get me for one excited (I’m very exciteable).

Kev

#6

I will grab some screen shots for tomorrow…this is all I have at the moment

http://home.mindspring.com/~hawkwind/xg4.jpg

#7

jnlp is quite simple actually. I do believe Mr. Kevglass came up with this page on it

http://www.cokeandcode.com/info/webstart-howto.html

Which helped me immensely in getting jnlp going.

As for webspace, what about java.net, sourceforge, javagamesfactory, etc? If you just want to throw something up there for the java-gaming.org community to see, I’ve got spare webspace/bandwidth I could give ya.

#8

Looks like a great example of a Xith game too :slight_smile:

Kev

#9

Looks really, really great. How will you port the collisions system part to the current release? Maybe you could use the HUD system instead of the SwingGUI (If you want, I could part this part for you). How many classes does this game have? Or in other words: Is it a big game from the view of a developer?

Marvin

#10

Bearing in mind we did Survivor in just weeks, in free time, using Xith, back when Xith was a lot less good, I think anything at pacman level is aiming way way way too low.

The biggest issues to making it look good with low effort IMHO are getting good textures and models, so I’d suggest picking something where you have already found some free models + textures (we got free textures for survivor just from looking around, and changed the level + model design etc to fit), and then making a simple shmup, either 1st person or 3rd.

#11

Waow I didn’t think this post was gonna cause so much feedback. I’ll try to answer intelligently.

Sorry man, I’ll think about it next time.

I agree. Sometime you take the same time to actually write a task than you would have spent to actually DO it.

Stratagem

  • 3D networked real-time massive strategy game
  • Open-source, Java, uses Xith3D^JOGL
  • Uses the Cal3D tech, Swing GUI
  • ENTIRELY “flavorable” by the mean of .xml files (in other words, you can change the entire content, including the units attributes which are countless, and everything else)

Enjoy : http://mkjeu.free.fr/Stratagem-MakingOf/Screenshots/
Screenshots are not all actual (look at them in chronological order or you will have surprises).
I’m currently fighting with wrong mouse events (Swing gui problem, I think), it’ll be fixed soon (I hope).
Next I’ll rewrite IA, taking into account every attribute we’ve designed (the design board). Then I’ll rewrite network support (probably reworking data structures which get serialized or just abandon serialization and get back to byte-per-byte control).

I’m an user, I’m making a game, I’m a painful user, It happens to me to fix bugs, to answer noobs, to implement features, to write docs, and so on.

Well, jME is something you didn’t know and the unknown is always interesting, isn’t it ?

It’s going to be the most developer-oriented way to create games, be it open-source or commercial.
I think we should emphasize on amazing graphical debug features (including mechanisms to debug Xith3D itself) and runtime adjusting feature, maybe include a shell (DynamicJava ? http://koala.ilog.fr/djava/) would be just wonderful.
We have to do everything we can to simplify our developer lives ! Let your lazy side come out (talking about ideas, not actually working on Xith3D).
Xith3D is a tool, and I’m gonna modify it in every way I need to (trying to keep the others happy). During a long period of time I wondered how far can I go in contributing to Xith3D now I know it’s ready to be taken, I know we can take control over it and make it dreamlike. We shall find things which differentiate us from jME. Instead of trying to have at least the same features, let’s develop what we need.
Sorry if you already knew all of the preceding, I said it more for me than for the others. If you have the same dreams as me let’s make Xith3D better together ! Go for it !TM

#12

See you’ve done it again. Don’t assume. I used JME alot when it was early days too. The first version of the space hulk game that I’m re-attempting now was written using it: http://www.cokeandcode.com/node/304

So it’s not new. Given I’ve used Java3D, Xith3D and JME extensively in the past I’d consider my choice pretty informed - unlike some evaluations.

Looks interesting (at least from the number of models pov), but the screenshots so far don’t exactly inspire (me at least). However, hawkwind seems to have something pretty exciting looking - so hopefully that’ll get public playable nice and soon.

Either way, my personal opinion really should mean squat to you if you’re getting your game out then it’s got to be all good.

Have you ever considered a Xith Spotlight on the webpage? Showcase your game screenshots, hawkwind’s and whatever everyone else is working on. I notice the JME showcase gets people pretty interested pretty quick.

Kev

#13

[quote="<MagicSpark.org [ BlueSky ]>,post:11,topic:28308"]
Enjoy : http://mkjeu.free.fr/Stratagem-MakingOf/Screenshots/
[/quote]
I’m really very impressed. Good work.

Amos and me were once talking about a place to present your game and to collect some games written using xith. I think we should make xith.org this place. When I’ve proceeded working on my game (an ascendancy/“master of orion” clone), I’ll put it there as well.

Marvin

#14

[quote]Looks really, really great. How will you port the collisions system part to the current release? Maybe you could use the HUD system instead of the SwingGUI (If you want, I could part this part for you). How many classes does this game have? Or in other words: Is it a big game from the view of a developer?

Marvin
[/quote]
1 - I definately want to port to the new HUD system. I wrote a “theme-able” hierarchy which is basically a background panel containing a top panel (Icon / Title bar), a bottom panel (L/R/ buttons and exit button), and a container central panel for game/dialog contents. Each GUI element takes a constructor arg that is a class containing a set of images to theme the window. By manipulating the opaqueness of the various components I get different effects (no buttons is achived by setting opaque to false and leaving the panel empty for example… In my personal shame I have many, slightly different, dialog windows. These differ mainly in text content…but…my version Xith didn’t have Text2D working. For me to easily use the HUD I would need to be able to layout a layered window. From what I have seen of your demo’s this is certainly feasable.

2 - I have a resonable number of classes which are on the small side. I really need to perform some cleanup because the debris of many failed ideas is contained there. Certainly my entrie GUI system will be replaced using Text2D and HUD stuff.

3 - On another note my scenery is generated with “Art of Illusion”, a open source, java based, 3D editor. I have added plugins to export the scene objects along with additional info telling me Object X is located at position x,y,z. or Light A is at x,y,z with color RGB. Other attributes including hidden/visible, game type, pickable, speech, etc. While this is NOT Xith specific, it might be interesting to generalize export Data such that a new Xith loader could pull and entirer scene with gaming attributes. This would/could be the first version of a Xith GUI editor…just a thought.

4 - Collisions…Basically in the old Xith system you added N objects the the Collision system, and use a Force, and it works. Odejava seemd to have a Xith<–>Object conversion to simplify this. Odejava site is down so I cannot verify. Handling the collisions is isolated so if the JOODE project has some examples I can most likely adapt.

BLAHBLAHBLAH…Dude I am so tring to get my stuff into javagamesfactory…it never seems to load my jar. Basically I have my jar, some old Xith jars, and the jar set associated with the freetts speech synthesis project. ANy help appreciated.

#15

[quote]Waow I didn’t think this post was gonna cause so much feedback. I’ll try to answer intelligently.
[/quote]
Way cool…Stratagem

#16

Do you know, that the HUD system is themable from itself. Any Widget class has a public inner class called Description, which does exactly what you describe: It holds all common information for a Widget type (like Textures). This can also be handled all together by a theme. Have a look at org.xith3d.ui.hud.resources.themes.

Marvin

EDIT: Feel free to ask me how to use your theme, when you’ve created one. And I’ll really love to see the first complete theme created after my own (the default one) :wink:

#17

http://www.xith.org/news/want_your_game_to_be_known_.html

Current slogan : Xith3D, Go for it ! TM

#18

Wow! The Xith3D website is instantly more enticing!!! :slight_smile:

Kev

#19

Okay, now I know the way to go ;D But we shouldn’t polish the website and forget the engine ^^

#20

Absolutey, but then if you focus on your game (which we be helped hopefully by people getting excited by elephant screenshots! :)) then that should drive the engine development and usability focus along nicely.

Thats an interesting one - could be pretty damn nice.

Kev