more angle problems

ok this one seems to be a tough one. Well for me anyway :)

is it possible to use angles to determine whether an object has passed another object. Have a look at this

http://img318.imageshack.us/img318/9014/problemwu4.png


you see that this cannot work, also checking the axis seperatly wont work either. Because the same problem can occur, also note that gravity can effect the object, so the object can be released parallel to the center and might fall below the center before it visually passes the center

any help?

ok well this works, but the object must not be affected by gravity until it has passed, meaning its trajectory stays the same

//rd = distance vector - floored
//rel = inital vector - floored

var passX = rd.x >= 0 && rel.x <= 0 || rd.x <= 0 && rel.x >= 0 || rd.x == rel.x;
var passY = rd.y >= 0 && rel.y <= 0 || rd.y <= 0 && rel.y >= 0 || rd.y == rel.y;

if(passY && passX) //reset gravity

if you use localtoglobal to get the red balls position relative to the white ball then the red ball has passed through the white ball (or it's axis) when the sign of the X or Y has changed.

Just thinking out loud here, but maybe you could "rotate" the x axis, so that the ball is at 0,0 and the moving ball is along it, if it goes to negative it crossed.

(The x-axis would be along the actual line drawn between, not the direction of movement)

i might try that out blink, UnkownGuy care to elaborate, i kinda got lost

Hm, can't you use the dot product of the velocity vector and the distance vector between the two object ( from the red ball towards the center ). If it's negative
it will have passed it.
Hope I didn't misunderstand the question?

I don't mean to be the simpleton but it seems like this problem can be resolved with a hitTest. I don't see any reason you can't.

i also had a kid ask me about the accuracy of 2 points. I moded the code some. I may not be what your looking forbut I think it will give you the idea.

Code:

---

range = 40;
closeness = 0;
function onEnterFrame() {
        pointcode();
}
function pointcode() {
//this just makes the hero move for the test
hero._x += 10;
//checks to see if its in the radius well just the X
        closeness = checkpoint._x-hero._x;
        if ((arrow_._x>checkpoint._x-range) && (hero._x<checkpoint._x+range)) {
               
                trace("good");
        }
//checks to see a miss
        if (hero._x>checkpoint._x+range) {
               
                trace("missed");
        }
}

---

you could add the y to get the all around checking. You can make the range 0 to be super accurate. This code was used to help make a DDR game so I think its good, being that DDR is all about checking checkpoints .

trying to grasp at what you're asking, you want to be able to do a hitTest succesfully against a bullet even when it travels really fast? i thought tonypa's covered this somewhere, oh well, let me think...

you ever heard of slope-intercept? you should have

Quote:

---

slope=(dy-ry)/(dx-rx);
yintercept=dy-slope*dx; //i think?
if(Math.abs(yintercept)<=combinedwidths){
if(dx/Math.abs(dx)!=rx/Math.abs(rx)){
trace("pass?");
}
}

---

i think that'd be good enough, unless ive gotten it wrong somewhere, my brain's gottin' all fogg-ed upp-ed

btw malee, you're angle thing could actually be used for guard ai, if the angle difference is great enough, they would be like, hey!! very nice, they'd react to slow close objects or fast objects that were far away

Quote:

---

Originally Posted by Kianis

Hm, can't you use the dot product of the velocity vector and the distance vector between the two object ( from the red ball towards the center ). If it's negative
it will have passed it.
Hope I didn't misunderstand the question?

---

thats exactly it, thanks Kianis works a treat

Quote:

---

Originally Posted by mr_malee

thats exactly it, thanks Kianis works a treat

---

Great :D Glad I could help