gametechmods
Off-Topic => Games => Game Development => Topic started by: System32 on September 28, 2009, 10:56:15 AM
-
Find Object name
Find Folders
If Folder = Object name
Find Sprite
Find GMF
Find Points
Combine Sprite to match Points X,Y Variables
Find Data
Import Data
Hitpoints.Data = hitpoints
Damage.data = damage
End If
Anyone wanna help with it? I Also think we could Use Box2D for physics..
-
Box2D is well... 2D.
I still can't get myself to start working on a graphics engine. And I still have no idea for the physics engine, I'm afraid I'll have to write my own, too. Or modify Bullet / ODE a lot to include some basic soft body physics.
System32, what language(s) do you program in?
-
for the graphix engine opengl would be perfect(i think)
-
for the graphix engine opengl would be perfect(i think)
OpenGL is not a graphics engine, it's a specification that graphics card drivers (are supposed to) follow. Direct3D offers an analogous functionality. And yes, between those two, OpenGL is more likely to be chosen.
-
ok then code itw/ c++
for the models use some file format that anybody can easily edit, like .blend.
-
ok then code itw/ c++
for the models use some file format that anybody can easily edit, like .blend.
**** Blender.
I was thinking of COLLADA.
-
Why? Blender is awesome, and free.
-
free means anybody can use it.
of course blender has .collada importing and exporting, but it sucks.
-
Box2D is well... 2D.
I still can't get myself to start working on a graphics engine. And I still have no idea for the physics engine, I'm afraid I'll have to write my own, too. Or modify Bullet / ODE a lot to include some basic soft body physics.
System32, what language(s) do you program in?
VisualBasic, English, French, HTML and I'm learning Python and LUA in my freetime.
I'm saying Box 2D because I feel it will be Infinately Easier to Make it 2D first,then work up to 3D.
-
definately, but it will be a top down type thing and the only things that world work well for phisics are horizontal spinners, rammers and pushy type bots. the z dimension fro vertical stuff would be hard to implement in a 2d environment.
-
Sorry, but I don't think so. Scaling up from Box2D and / or a 2D graphics library (like SDL a.k.a. pygame) to a full-fledged 3D game engine is just going to be a massive rewrite. Everthing starting from basic game class, through algorithms up to file formats will have to be rewritten.
I was rather saying that using blender's file format for a game is plain retarded. Also, Blender really does suck. I hate it with a passion, just like The Gimp. The user interface (or rather the lack of it) is a nightmare. Good things must cost sometimes.
From what I saw collada is a great file format, since it can be (re)used eveywhere, it's open source, and easily parsable.
I was taking a look at irrlicht, but I think it may be a little bit overkill for our purpose. But it being written in C++ and released under the zlib license gives it a huge plus.
System, learn python, do some simpel projects, like a soft body physics engine (using verlet integration, http://en.wikipedia.org/wiki/Verlet_integration , wikipedia explains it well), then use it to create some fun physics based games, add soem collision support, etc. It will teach you all the basics of game developement. Unfortunately VB is a no-go for this project - keeping it platform-independant is a must. Don't get me started on Mono.
-
Unity Is a useful thing.
http://unity3d.com/unity-web-player-2.x.html
Free, Safe Download... And It's 3D!
But I'll start with Python....
-
Unity Is a useful thing.
http://unity3d.com/unity-web-player-2.x.html
Free, Safe Download... And It's 3D!
But I'll start with Python....
Why run crap slowly in the browser when you can run it lighting fast natively?
-
thats why you export it in .exe and not webplayer.
i have and use unity all the time.
-
My Grapics card is OLD!
My PC is old!
It runs RA2 perfectly so I don't need the Nivida patch.
Unity runs smotther than enything else for me.
Edit: Screw that! 999 to Make stuff ith it!
-
*sigh*
Why won't you just goddamn get it? A RA2-like clone is unfeasable in existing engines because of how specific it's geometry and physics handling is. One way or the other there will be a necissity of writing a large amount of low-level, boring code.
Dynamic geometry and physics proxy generation and modification (necessary for the CREATE part in the game's idea, as well as the DESTROY part) is nowhere to be seen in ANY framework on the market.
-
A RA2-like clone is unfeasable in existing engines because of how specific it's geometry and physics handling is.
Somone said that it traces a certain object Constantly per ping. (You know, It keeps checking up on where it is and where it might go every ping, and changes what the collisions may be accordingly.)
So, Could We add a code that Connects The Attachment points to a certain way, and when said object spins it isn't going to be checked because the object is attached... Like you won't check up on what a tyre does because you know it's fixed to the motor, and you only need to check what damage it has.
Dynamic geometry and physics proxy generation and modification (necessary for the CREATE part in the game's idea, as well as the DESTROY part) is nowhere to be seen in ANY framework on the market.
Kinda hard to understand. Can you explain?
-
Imagine a simple game engine, like Ogre3D. The way you set up a scene is:
1) Preload your sh** (3d models)
2) Position your sh**
3) Get your input
4) Animate your scene
5) Go to 3).
Where 4) (Animation) can only do stuff like movement and rotation, not modify the geometry of an object.
Imagine your bot gets it's chassis bashed up. Uh-oh, gemotery can't be changed.
Even worse if you get a ready-made physics binding (like Ageia) for your standard game engine. Chances are it doesn't support independant wheels.
It's hard to explain. I tried, I failed, people on the forum of Ogre3D said the engine is not made for this type of games.
-
you could have a massive phisics engine that only quad core 3 ghz computers would be able to do... remember my ideas for ra3 serge?
-
not modify the geometry of an object.
Softbody physics?
-
That's totally unnecessary.
I think the only way a physics engine could handle this sort of stuff is by using a node mesh to represent how strong is a component. Example:
(http://imgur.com/rDWr6.png)
In algorithm terms, the nod emesh calculation is really simple to do. Incorporating soft body physics with rigid body physics gives us both speed and some form of distortion. That's why no stock stuff will do.
-
Softbody physics?
Pure softbody is a bit slow, and AFAIK there is no open-source engine made for game developement, and writing one from scratch would be a nightmare, especially collision handling and so on. That's why combining the two gives us the best of both worlds - fast and stable rigid body physics + minimal soft body physics for component distortion.
-
good idea!
what about the components sticking together, would u use my idea of have them actually interlocking or have some connection point?
-
Pure softbody is a bit slow, and AFAIK there is no open-source engine made for game developement
what about Dark Basic ? it is quick to learn, fast feed back. it propose some collision management etc.
-
All can be done using these node meshes, so that one node mesh is used for the whole robot, and components don't alway fall of at their attachement point, but at the weakest part. The nodes and connectors can have a different strength attribute to differentiate how strong a part of the robot is.
-
Serge, thank you for explaining that.
-
what about Dark Basic ? it is quick to learn, fast feed back. it propose some collision management etc.
Why darkbasic, when there is C++, eventually C++ using python for scripted stuff, like RA2. Python is easy enough for everybody to learn, and yet is not proprietary.
-
Serge, thank you for explaining that.
You're welcome.
-
Hmm. I thiink I remember a Physics engine for that.....
Aha! Endorphin.
Not saying we should use it, just remembering.
Also, Did you hear me out on my Idea to cut down on the cause of LAG on RA2?
-
that would work for wheels, but not for spinning weapons...
-
Gah!
Get out.
That I'm saying is By having an attachment point, and nodes around it, You can just keep an object fixed to another object while still having it move!
With that, You'll be able to have, lets say... a tribar with 3 spikes, and have is spinning while going forward, because the Iron's are Fixed onto the tribar unless it breaks, and the tribar fixed on to the Axle, which in turn is fixed (But still allowed to turn under a control) within the motor, which is what holds the axle!
Attach two pieces of Lego and shake it.
See? It won't fly off, so you Don't need to focus on it unless you hit it off something.
Combine that with the Node system serge talked about and some basic physics and we have out collision system sorted! then we can work on the Movement system, repair calculations and as a final possibility, an improvement on the modifiability that RA2 has given us! (AKA, AIing is made easyer, Components can be made using Txt and BMP files, stuff like that.)
-
ohhhh i thought u ment that it wouldnt have to work out the phisixs for them.
-
It wouldn't! well, the part that causes lag online wouldn't.
-
there will still be some lag... the only way i can think of a way to getting around this is to have dedicated servers and some very good tunneling.
edit: gah why cant atari just release the source! its not like they make much money off of it and they can still sell the other part of it(like the quake and doom series is doing.
-
Somone said that it traces every object Constantly per ping.
THIS IS WHY IT LAGS!
THE MOTHER****ING CODE CHECKS UP ON EVEY ****ING COMPONENT WAAAY TO MUCH, THUS IT LAGS ONLINE!
I'M SAYING WE SHOULD REMOVE THAT BIT OF CODE!
GOT IT!
-
ooooookaaaaaaayyyyyyyyyyyy mr scary sweary
-
that would work for wheels, but not for spinning weapons...
Components aren't going to be "bouncy". The constraints will just stretch when enough force is applied.
-
Wait....So what you are all saying that there's no other source code or phisics engine, that at least mimics RA2?
Damn...Well might as well ask atari if they won't sell and start some sort of donation...Methinks that enough cash will convince atari to sell the rights to the engine.....(Kinda kidding myself here)
But that still seems a bit imposible, I know it would have to be something that runs on older machines as well, but aren't there any physics engines that enable dents and stuff like that? I mean come on Worms 4 has destructible terrain and that's 3D :|
Or is it more complex than that.
Also about the mesh idea.
That would be painstaking to write, the code i mean.
But still if there are no other options?
It's kinda time for a new robot destruction game, and i don't mean arcade fighting.....
-
Wait....So what you are all saying that there's no other source code or phisics engine, that at least mimics RA2?
Well, of course you can use a rigid body physics engine, but that'll just be RA2.5.
Damn...Well might as well ask atari if they won't sell and start some sort of donation...Methinks that enough cash will convince atari to sell the rights to the engine.....(Kinda kidding myself here)
We tried to get the source code from them. Impossible, since it uses Havok, which has an expensive license.
But that still seems a bit imposible, I know it would have to be something that runs on older machines as well, but aren't there any physics engines that enable dents and stuff like that? I mean come on Worms 4 has destructible terrain and that's 3D :|
AFAIK there aren't any that aren't proprietary. Also, Worms 4 doesn't have a "real" physics engine, just terrain destruction (no complex collision detection).
It's kinda time for a new robot destruction game, and i don't mean arcade fighting.....
It is, but writing a 3d game is a lot of work. I am still working on a graphics engine, however right now I'm on holiday and don't have access to my development box.
-
While we're on this subject of engines, I might as well provide you with one that caught my attention.
http://en.wikipedia.org/wiki/Digital_Molecular_Matter (http://en.wikipedia.org/wiki/Digital_Molecular_Matter)
What do the pros think?
-
While we're on this subject of engines, I might as well provide you with one that caught my attention.
http://en.wikipedia.org/wiki/Digital_Molecular_Matter (http://en.wikipedia.org/wiki/Digital_Molecular_Matter)
What do the pros think?
It's awesome, but probably very expensive.
EDIT: Apparently they're releasing a free API soon. w00t!