yes, but this isn't the problem...

the problem is that if you keep moving against a wall the ball begins to spin and drills through the level geometry.

but in my level it's almost impossible that it happens with the current settings... (it was very tricky to find these settings though!)

Nice demo ventilator. I couldn't get the ball to fall through the level after trying for a couple minutes. It is a trivial issue if it takes a lot of work to make it mess up.

I was thinking of trying to use trace as a method to limit the problems I had earlier. I hadn't quite thought of how I wanted to approach it though. I like your movement through torque as well.

I'm not sure of how to limit that odd spin reaction either. Can you think of something that would reduce/eliminate that? Or why that might occur at all?

thanks,
Jason

I've gone through your code some more ventilator. It's very insightful.

Removed questions and moved to ask conitec forum.

thanks,
Jason

it's just my style. i don't know if it's faster.

code:

---

vec_set(dir,nullvector);
if(key_a)
{
  dir.x-=1;
}
if(key_d)
{
  dir.x+=1;
}

---

the reason i did it this way is that when the user presses a and d at the same time dir.x should be 0. i think it's more elegant than with an additional if statement.

btw. i forgot to time correct the force applied to the ball. the vec_normalize line should be something like that:
vec_normalize(dir,50*time);

Seeing your control scheme helped me realize that the way I was doing it was incorrect. I didn't realize I was messing up the actual input in the manner I had it scripted. The method that you have shown is much better. The code is cleaner and logically correct.

I am curious about the speed of some script commands over others. I'll ask the script questions on the ask conitec forum.

Thanks again for your help. It's made it a lot easier for me to realize how much I need to improve my coding techniques.

Jason

Are you referring to the ball falling through the level boundaries (skybox) or through blocks inside the skybox ? I could only reproduce the ball falling through the skybox.

I was referring to the ball falling through level blocks, not the sky box. I haven't had any problems since reducing gravity and moving the ball with both torque and forces. However, when I moved the ball with just a force and no torque it would start spinning and go through the level blocks if you applied the constant force on the ball somewhat perpendicular to the level blocks(Or if you were facing a corner).

Jason

marco, in my example the problem isn't noticeable anymore with the current erp/cfm settings (although if i really try hard i still can let the ball fall through level geometry).

i first thought jason did apply too high forces but i easily could reproduce it in my levels. if a ball is forced against a wall it starts to spin faster and faster (it doesn't matter if you use a force or a torque to move the ball) and then it moves inside the level geometry.

yes, i recovered the same problem in with my metallballs! they were jumping from wall to wall spinning faster and then dissapearing through the wall

I noticed this twice with the old physic beta.
Later with a newer version never again.