Want to Do: Thinking of writing a new tutorial, what would you guys like to see?

[zuperxtreme]

I'm bored and have some time on my hands, what sort of game would you like to see a tutorial of? Tower Defense? Some sort of RPG? Physics? Let me know (2D only).

[is all mememememe with me]

this is more of an irl game rather than a computer game, but...

how to kill darkhunter in the most painful way possible

[HampTheToker]

I want to see a tutorial for Slan Bag, The Game.

On Topic: Do you do Android games, perchance?

[zuperxtreme]

Nope, but I really should. Only Flash games for now. The logic should be simple to port to something else, it's all of the same.

[HampTheToker]

Relevant: http://www.fxitech.com/products/

Will make coding Android games and apps much easier for fledgling Android developers.

[Fuck "Fuck Fuck" Fuck]

physics AND tower defence, combined

[zuperxtreme]

Got anything in mind? Using Box2D isn't TOO hard.

[Lanny]

Tetris

[Qhost]

A AS3 tutorial for certain RPG mechanisms. Such as an inventory with some advanced features (click and drag around/swap, automatic sorting, stacking etc). A NPC interaction system would be great (speech, buying from a shop). Also some form of save system to save all of this.

Those are the 3 tricky areas I always have when looking for tutorials for RPG's, esp. in AS3...

[zuperxtreme]

So a physics, Tower Defense, Tetris mix with RPG elements?

What could go wrong?

[Lanny]

Heh, well I'll withdraw tetris and switch over to Qhosts suggestion. I'm sure it can't be that hard to learn how to write a tetris clone. A graphic RPG would probably be more interesting.

[Valheru]

I have a challenge for you. On a 2D plane, have a cannon, shooting a projectile with a fixed, known velocity, calculate the lead angle correctly to hit a moving target.

The vector of the moving target is known to the cannon (assume it has radar or sumshit). The projectile has a constant velocity, straight line only.

If you figure it, you'll definititely have my respect. It's not an easy problem to solve.

[Lanny]

Originally Posted by Valheru

I have a challenge for you. On a 2D plane, have a cannon, shooting a projectile with a fixed, known velocity, calculate the lead angle correctly to hit a moving target.

The vector of the moving target is known to the cannon (assume it has radar or sumshit). The projectile has a constant velocity, straight line only.

If you figure it, you'll definititely have my respect. It's not an easy problem to solve.

Does the target move in any direction, or just perpendicular to the cannon (like can it move away from the firing position)?

[Valheru]

it can move in any direction. Assume the cannon is a dot mid-screen, and the target is a ball bouncing off the edges of the screen in straight lines (think dxdiag). The cannon should be able to calculate at any time in which direction to aim to ensure the projectile hits the target.

This problem is gonna fuck with you. It requires a neat bit of math-fu.

Ps. Ignore the requirement to hit if the ball target bounces off the screen edge after the projectile is fired, obviously it will miss (but would have hit if the edge wasn't there, capisci?).

And fucking lasers aren't allowed!

[Lanny]

Either way you could do it simply (although probably not elegantly) by calculating the path of the target to some resolution (say 10 points per second along its path, that would define your margin of error) and index them by distance from your firing position. Then calculate the distance from your firing position your "cannon ball" would be at the same resolution. Take the time that the closest match occurs, figure the position of your target at that time, and fire at that point.

[Lanny]

Math-fu? Bah, processing time is cheap, easier to approximate it to a high resolution

[Valheru]

Yeah, but ammo ain't. Do it right and don't be a bitch.

[Lanny]

But you don't need more ammo with my approximation than whatever other calculation you could use .

[zuperxtreme]

It's not so cheap if you're calculating against a hundred, or a thousand objects. In Flash at least.

My ghetto way would be to just shoot missiles that follow the enemy, not bullets.

To actually hit the enemy with bullets a quadratic equation would help: http://gamedev.stackexchange.com/que...et-to-an-enemy

No I don't know it off the top of my head. >_>

[Ethan]

How to build the core functionality of a roguelike from scratch. Next? Pathfinding AI.

[Valheru]

Originally Posted by zuperxtreme

It's not so cheap if you're calculating against a hundred, or a thousand objects. In Flash at least.

absolutely. I want to see a brute force method give playable framerates in a game like supcom.

My ghetto way would be to just shoot missiles that follow the enemy, not bullets.

Variety is the spice of life. Besides, bullets can't be countermeasured . And to have an efficient intercepting missile that maximises its range, the missile also needs this algorithm.

To actually hit the enemy with bullets a quadratic equation would help: http://gamedev.stackexchange.com/que...et-to-an-enemy

No I don't know it off the top of my head. >_>

100% correct. The trick is quite counterintuitive, you have to calculate how long the projectile would take to hit, if it hits. Only way as far as I'm aware is to calculate time t as the roots of that quadratic, giving you a positive and a negative value (which you discard). Given t, you can now use standard trig and displacement math to work out the lead angle.

Doing it in 3d is a trivial matter, you just have to account for the extra dimension.

[Lanny]

Interesting! I still secretly like my brute force method, but honestly it's kinda shit. I was bored though so I ended coding it up. Apparently python has a little module that's perfect for this kind of thing.

#!/usr/bin/python

# Lanny's targeting aproximation script

from turtle import *
from time import clock
import math, random, time

MARGIN_OF_ERROR = 3.0
TIME_TO_SLEEP = 0.01

class Projectile(Turtle) :
    def __init__(self, direction, startspeed, startx, starty) :
        Turtle.__init__(self)

        # Storing this here just to avoid reverse calculating later
        self.startspeed = startspeed

        self.pu()
        self.setpos(startx, starty)
        self.pd()

        self.setheading(direction)
        
        alpha = self.heading() * (math.pi / 180.0)

        self.vx = startspeed * math.cos(alpha)
        self.vy = startspeed * math.sin(alpha)


    def step(self) :
        x,y=self.pos()

        self.setpos(x+self.vx, y+self.vy)

    def start(self):
        step()

def FiringSolution(target, speed) :
    '''Solve for a trajectory based on target properties and speed'''
    MAX_CATCH_UP_TIME = math.sqrt(target.pos()[0]**2 + target.pos()[1]**2) / speed - target.startspeed
    TargetLocationByTick = []
    for tick in range(int(MAX_CATCH_UP_TIME)+100) :
        x,y = target.pos()
        
        x += target.vx * tick
        y += target.vy * tick

        TargetLocationByTick.append((x,y))

    TargetLocationByDistance = []
    for loc in TargetLocationByTick :
        TargetLocationByDistance.append(math.sqrt(loc[0]**2 + loc[1]**2))

    print str(TargetLocationByTick)

    CurrentOptimumInterceptTick = 0
    CurrentOptimumInterceptDistance = 99
    for i,dist in enumerate(TargetLocationByDistance) :
        if fabs(i*speed - TargetLocationByDistance[i]) < CurrentOptimumInterceptDistance :
            CurrentOptimumInterceptDistance = fabs(i*speed - TargetLocationByDistance[i]) 
            print CurrentOptimumInterceptDistance 
            CurrentOptimumInterceptTick = i

    TarLoc = TargetLocationByTick[CurrentOptimumInterceptTick]
    print 'TarLoc:' + str(TarLoc)
    OptimumInterceptTrajectory = math.acos(TarLoc[0] / math.sqrt(TarLoc[0]**2 + TarLoc[1]**2)) * 180/math.pi

    return OptimumInterceptTrajectory



if __name__ == '__main__' :
    Target = Projectile(random.randint(0,360), random.random()*2, random.randint(0,100), random.randint(0,100))

    Bullet = Projectile(FiringSolution(Target, 3), 3, 0, 0)

    while True :
        Target.step()
        Bullet.step()

        Bx,By = Bullet.pos()
        Tx,Ty = Target.pos()
        if math.sqrt((Bx - Tx)**2 + (By - Ty)**2) <= MARGIN_OF_ERROR :
            print 'INTERCEPTION COMPLETE, YAY!'
            break
        time.sleep(TIME_TO_SLEEP)

[Valheru]

Nice. You have the gift.

[zuperxtreme]

Lanny, you should learn Actionscript so we can collaborate and make something.

[Valheru]

Lanny's firing solution, err, solution, is actually very clever. It may not be optimal, but it's typical of clever programming hacks where a mathematical solution is not readily apparent, or possible. It's basically a hill-climbing attack, and it never even occured to me. Some things can only be solved this way, maybe with some basic heuristics thrown into the mix in order to improve efficiency.

[Lanny]

Originally Posted by zuperxtreme

Lanny, you should learn Actionscript so we can collaborate and make something.

I'd love to! What ever happened to those tutorials you did? Do you know anything like "Actionscript for programmers"? Like something that's more of a "here's the syntax and how AS varies from other languages" as opposed to "here's how to program".

Originally Posted by Valheru

Lanny's firing solution, err, solution, is actually very clever. It may not be optimal, but it's typical of clever programming hacks where a mathematical solution is not readily apparent, or possible. It's basically a hill-climbing attack, and it never even occured to me. Some things can only be solved this way, maybe with some basic heuristics thrown into the mix in order to improve efficiency.

Aww, that's one of the nicer things someone has said to me. Thanks .

[zuperxtreme]

Originally Posted by Lanny

I'd love to! What ever happened to those tutorials you did? Do you know anything like "Actionscript for programmers"? Like something that's more of a "here's the syntax and how AS varies from other languages" as opposed to "here's how to program".

Depends, coming from what background?

Here's one coming from Java: http://www.infoq.com/articles/actionscript-java

My tutorials are outdated(there's a new Flixel version), so you'll get a bunch of compile errors.

You'll need FlashDevelop: http://www.flashdevelop.org/download...-4.0.0-RTM.exe
The Open Source Flex SDK: http://opensource.adobe.com/wiki/dis...3201&pkgtype=2
Flash Player 10.1 ActiveX runtime, this one: http://fpdownload.macromedia.com/pub...ebug_32bit.exe
And this one(I think): http://fpdownload.macromedia.com/pub...ebug_32bit.exe
Flixel: https://github.com/AdamAtomic/flixel/zipball/master
(and then Java 1.6+, .Net, blah blah)

If you're interested, Flixel has a bunch of Demos with the Source you can look over: http://flixel.org/features.html

Or PM me and I'll help you out.

[Wolf]

How about a top-down shooter like the old GTAs?

If you really feel like it, then maybe even a multiplayer version.

[Lanny]

Isometric style game? I'm doing something similar in python (because I'm lazy and don't want to start a larger project in a language I don't know yet ) and I'd be interested in seeing how other people would do it.

[zuperxtreme]

From scratch or are frameworks allowed? ;O

[Lanny]

Either is fine, it's your tutorial after all . From scratch might be more interesting, but would probably take an inordinate amount of time.

[zuperxtreme]

Pew pew http://www.swfcabin.com/open/1330542993

I was starting on that, so maybe I'll do a tutorial for something like that. Idk, been really lazy lately. :|

[Wolf]

Originally Posted by zuperxtreme

Pew pew http://www.swfcabin.com/open/1330542993

I was starting on that, so maybe I'll do a tutorial for something like that. Idk, been really lazy lately. :|

Aactually, a tutorial with a decent AI (pathfinding etc) would also be awesome. I've been making flash games for a while now, but AI is one of my weak spots.

[L33tz]

java game please.

no flash nonsense

[Lanny]

Java as a web platform is dying, and with good reason too. Not that Flash is ideal either, HTML is the way of the future, but seeing how fast adoption is happening, I'd rather get to know Flash than Java.

[Wolf]

Originally Posted by Lanny

Java as a web platform is dying, and with good reason too. Not that Flash is ideal either, HTML is the way of the future, but seeing how fast adoption is happening, I'd rather get to know Flash than Java.

AFAIK, HTML5 will replace Flash sooner than any kind of HTML will replace Java. But that's just based on the opinions I've heard.

[Lanny]

Originally Posted by Wolf

AFAIK, HTML5 will replace Flash sooner than any kind of HTML will replace Java. But that's just based on the opinions I've heard.

Does Java do things that HTML5 and Flash can't, and probably never will? Yeah, I'll give you that, but it still has not place in webdev to start with. I mean, yes, you can build "richer"(god I hate that word) applications in java, but the only time Ive seen Java used and actually needed in webdev is for browser based MMOs and stuff like that which shouldn't be browser based in the first place.

[L33tz]

Originally Posted by Lanny

Does Java do things that HTML5 and Flash can't, and probably never will? Yeah, I'll give you that, but it still has not place in webdev to start with. I mean, yes, you can build "richer"(god I hate that word) applications in java, but the only time Ive seen Java used and actually needed in webdev is for browser based MMOs and stuff like that which shouldn't be browser based in the first place.

wtf does this have to do with making games in java

[Lanny]

Originally Posted by L33tz

wtf does this have to do with making games in java

You shouldn't be doing it in the first place.