Its still the same just instead of code you use puzzle pieces and sliders and inputs to set parameters. But the logic is similar to running a script.

OOP style and patterns like composition ect. will be useful for this.

Just to let you know those guys were at least students (if not doing their phd) at MIT who developped scratch, so basically its a tough project for one 15 year old.

How about having an option pannel to change key mappings and let the user decide how to play?...

Or maybe you should actually look at assembler. Make a compiler that rolls out your scripts to flat code that can be run sequencially and uses jump tags and small basic atomic instructions. The complex stuff builds up on that basic stuff.

You won't get much further then doing a scripting language like this, may be you should find out what a compiler does and write an interpreter for your language. Then you will want to have some sort of interpreter/virtual machine thingy for it.

Check out LUA and similar script languages that already exist, specifically made for games. You can not just code out into the blue for this but have to plan every detail in advance.

At 11/20/09 03:56 AM, CloudEater wrote:

At 11/20/09 03:50 AM, hesselbom wrote: Nothing is as important as supper!

Are you sure!! Isn't breakfast more important then supper?

Some people don't eat breakfast... so..

Simply do the check that you doing at your characters center down at his feet. Like add a radius or offset vector.

At 11/19/09 12:43 PM, Deathcon7 wrote: This is not happening. Give up hope. It was a bad idea to begin with.

Well it was obvious that this would not go anywhere. Since making a good RPG (talking diablo, zelda, gothic, oblivsion anything real and no kiddy flash, not speaking of the 3d graphics side either) alone is already a huge task and then even wanting to make a framework or api for that is even more work. It would be possible if experienced people worked on it (not just one alone) and they had enough time. But trying this alone is madness :D

So don't bother asking for this to be released...

At 11/16/09 11:09 PM, Version2 wrote: That looks pretty amazing man, I wish I had your skill at coding. Gonna have a playable demo anytime soon? You can only get so much from a video :)

Thanks a lot! The engine does not have any real menus yet and the gameplay is not fully developed. As soon as i get some more time to work on it, i will try to put together some demo.

At 11/17/09 12:05 PM, Johnny wrote: I think its looking great so far. The lighting really makes it. I would work on the animation of the character himself a bit. He seems to 'float' while walking, instead of taking fluid steps. Everything is looking amazing though.

Thanks aswell!
The character animation is not final yet, so it might improve somewhat. Anyhow you probably mean the fact that his upper body does not bounce up and down with each step. Currently that will be tricky to do because of how the character animations are partitioned and animated. Once there is a good enough solution that convinces me, i might correct that.
Another thing is the feet being in the air when walking on slopes due to the player being a rectangle and not just a point at his feet. Like mentioned at the moment i dont have any specific solutions for these problems and rather want to nail the gameplay. So once the rest is fit everything will be improved as much necessary and possible. :)

Anyhow the engine has come a long way already and the road is winding and much longer ^^

Hey guys,

for those who don't remember, there is this small game project of mine, a sort of platform shooter.
Some of you might remember previous toying around with the convolution matrix ect. Anyhow there has been a huge improvement on the whole lighting model (thanks to the help of a friend, jazz). Basically the lighting can be considered more or less complete now.

Consisting of three major key parts, namely:
lighting & shadows
Normal Maps
Screen Space Ambient Occlusion

All together having a big impact on the graphics as you may see in this short preview video.

So how do you like it so far? Any suggestions or wishes for improvements or general feedback is welcome.

cheers

I will be away for a while so can't wait for you to post the code. But i just remembered having similar code in the key controller i had posted here a few days ago.

Check out this Link And have a look at the document class (Main.as) and KeyController.as (init(stage:Stage):void function). That is how it works.

Maybe it will help you seeing a working sample.

cheers

You are not passing the stage to the clip correctly and hence the referencing doesn't work. Your code doesn't show the source of the problem.

Please post the code of your document_class where you pass the stage reference to your class, and show the inner part of where you set the reference inside that class.

If you do it right it works so... the reference in the sample you posted has to be bugged.

hesselbom gave you a way to get it working.

Lets summarize 4 ways you can do it (including hesselborns):

- pass a reference to the stage in the constructor of your class to add

- have a set / add function which passes a reference to the stage and adds the class to the display list

- have a static variable that references the stage in your document class (then you can access it from anywhere)

- have a private static variable that references the stage in your document class and make a public static getter function

At 11/13/09 05:33 AM, hdxmike wrote: i think thats right

NO its horribly wrong in several ways, basically i wonder if you have any sense for logic after seing this code. Sadly its a mess with lots of unnecessary stuff that wont ever get checked... not to mention the unnecessary loop you do.

if(char.y+char.height<platform.y)
while(char.y+char.height>platform.y){
char.y= platform.y-char.height
}}

1.) The code will never run through the while loop!
2.) You don't need a loop at all.

This is what you wanted to write hdxmike:

if(char.y + char.height > platform.y){
// the char should be allowed to jump and only corrected when BELOW the threshold
char.y = platform.y - char.height;
}

;)

Maybe you should get a book on Linear Algrebra and Analysis, the math will always get back to you if you toy around with physics.
A physics book can be helpful if you don't understand the physics part of the equations either, although all you need to know is kinematics and dynamics.

Methanet is using separated axes theorem aswell, but there is a more general form of it that works for any polygon. The idea is to project one vector onto the other and do this for one out of two objects that could be colliding.

Once you know an object is colliding you will need to check at what point (usually only one or on an axes is the other case) and resolve the collision (hecker).

Next step is contact forces, and after that avoiding tunneling (sweeped bounding rects, sweep and prune: Barraff).

The stacking objects problem.

Then making objects sleep and wake up so collision detection and integration is only run when an object moves or collides.

Next stage would be optimizing your datastructure, spacial partitioning with a quad tree for example.

Once you passed all these stages you might have a working physics engine for rigid polygon bodies.

Next stage would be adding additional constraints... soft bodies... you can go on as long as you want and try to implement reality ... ^^

Maybe this gives you more of a summary how much effort it takes to get even a basic physics engine working. And you will need math mostly everywhere, as a solid programming knowledge foundation, since you have to work with algorithms, numerical problems and know your datastructures (ie. vectors, lists ect.).

The more realistic you want it to be, the harder it gets. Anything that involves rotation is already a lot harder to do then non rotated objects (unless their circles/spheres).

There are two things you could research and apply to solve the rigid body dynamics.

1. Collision detection by projection (rectangles or even polygons can be done using the same way). This also is in connection to the separating axes theorem -> if there are no separating axes two objects collide

2. Chris Heckers paper on rigid body dynamics (specially the impulse section, part 3 of the paper series).

You will also need a method of integrating stuff, here you can use runge kutta for high accuracy, sweeped bounding rectangles. Or you could do simple euler integration (will be buggy) or go with velocity verlets or verlet integration. BUT the toughest parts are 1. and 2.

And you can find plenty of papers/articles/tutorials ect. about all of these on the net, just by using google with the right search term. Most likely noone will help you code it though, so you are all by yourself :)

How about you post what you are doing so that one can actually see what is wrong?

You can check this thread for an alternative method: thread

Speaking of key controllers and OOP, there was this other thread about the same, let me summarize. :)

Usage:

import source.ui.keys.KeyAction;
import source.ui.keys.KeyManager;

// init keymanager
KeyManager.init(stage);

// bind some keys to the predefined gameactions
KeyManager.bind("UP", "w");
KeyManager.bind("UP", "UP_ARROW");

// in your enterframe / game loop
if (KeyAction.UP) {
	trace("UP");
}

And here is the whole thing: KeyController Source and Example

It uses Objects instead of arrays and you can bind more then one key to your specific game actions. All you have to do is instanciate as above and change the KeyAction class to your game's needs.

cheers

Here s an explanation that goes slightly more into detail then gusts link link

I made it a little more userfriendly and added comments, including example project for cs3 and flashdevelop 3.
It should be pretty easy to understand how it works, all you have to edit is KeyActions.as and bind keys to your games needs.

KeyController.rar

You can see that there are no loops necessary during runtime, only on initialization, to create the label/value lists.

Enjoy... ^^

Hm, I made one where you predefine game actions and then can map any key to any gameaction (multiple keys for an action ie.). Also you can remap keys ect.

Usage would be:

import source.ui.keys.KeyAction;
import source.ui.keys.KeyManager;

// init keymanager
KeyManager.init(stage);

// bind some keys to the predefined gameactions
KeyManager.bind(KeyAction.UP_VALUE, "w");
KeyManager.bind(KeyAction.UP_VALUE, "UP_ARROW");

// in your enterframe / game loop
if (KeyAction.UP) {
	trace("UP");
}

Downside is that you have to maintain a KeyAction file for your projects, but you will benefit from autocompletion and ease of use within your code.

I would not use globals, use a public class with static access methods instead, a "singleton".

At 11/6/09 02:03 AM, Argentin wrote: you just have to make a swimming animation and when the player touches the water, the character goes to the swimming animation frame and make the gravity lower and friction higher...

Physically seen, gravity stays the same!! Only friction is higher AND the water gives you an upstream (not sure if this is the right word in english).

So keep gravity, add force that works opposite to gravity, increase friction and you got water.

Check wikipedia on "polar coordinates"

@poster

You can look for tilebased tutorials for the squares and scrolling part. Then you need to know how to make buttons (code yourself or use flash buttons). The rest is just putting the pieces together. Try to break down problems into smaller parts and figure out solutions for those. Never go trying to find full tutorials on your specific and may be sometimes complex (in means of consisting of several smaller problems) problem.

If you dont want to use tiles (which seem perfekt for this) you can also just find out how to scroll something and how buttons work. And if you don't know yet, how functions classes ect. work.

At 11/3/09 02:40 PM, hdxmike wrote: DAMMIT ! forgot to say you need to add _root. to everyting ie
_root.map1.eufhshfjoa

Also i assumed you were using AS2

since everyone does :(

Your code isn't very useful anyways so don't bother. Anyhow i don't have the time to write down source either. This is actually done easily. Maybe the poster should ask some more specific questions on what part is causing him trouble. Noone will write the full code for you probably.

Specifically this one might interest you: Tutorial: Using Flash CS3 and Actionscript 3 to create Atari 7800 Asteroids

This guy has some great tutorials that might help you!

8Bitrocket.com

also have a look at
emanueleferonato.com/

Well in my engine i actually got rid of the z coordynate. Basically i am leaving out a component to only make stuff reflect in 2 dimentions. Not sure how to create the map for z components yet.

Anyhow the trick is that its a 2d engine trying to achieve effects used in 3d engines, which makes a lot of things different. There is no 3d data anywhere... the light is only working in 2 dimensions and also casting shadows in these and so do the normal maps. For visual reasons i highlight wall tiles instead of blacking them out even though the light would not reach them (speaking of surface on collision objects).

At 11/2/09 01:22 PM, Cryoma wrote: What are normals?

Generally a normal is a vector of length 1 or unit length. A normalized vector. Normal maps are graphics that represent the normal vectors on a surface of a texture. Each pixel color value stands for a component of the normal vector.

When light shines on a surface it gets reflected differently depending on the orientation. Normal maps are used to integrate this into flat 2d textured surfaces. In other words bump mapping. Correct me if i am wrong ^^