Transformed bounding box collision problem
@Shadowslair: Because that's a box oriented as the node and can be used for collision detection..... I thought that's what he was asking for....
@ChaiRuiPeng: bullet is indeed better than this, but this is probably a bit quicker
Edit: I do not mean quicker than bullet, I just mean this would be quicker to add to his program than including bullet
Sorry for my awful english ^_^'
-
- Posts: 363
- Joined: Thu Dec 16, 2010 8:50 pm
- Location: Somewhere in the clouds.. drinking pink lemonade and sunshine..
yes but its silly. its adding a whole other layer of abstraction. you might as well suggest he do a collision checking against every pair. or might as well suggest instead of using a shader that he make a billion little polygons and write a program inside a program inside a program to change those polygon colors in realtime to simulate a shader inside a simulation. all those ideas would work fine if we had an infnite resource on our computers but we have limited system resources. look at my signaturehayate wrote:
@Shadowslair: Because that's a box oriented as the node and can be used for collision detection..... I thought that's what he was asking for....
@ChaiRuiPeng: bullet is indeed better than this, but this is probably a bit quicker
Edit: I do not mean quicker than bullet, I just mean this would be quicker to add to his program than including bullet
in the long term it would cost much more to maintain. since bullet is already built for that. to maintain such system as you suggesting would be harder to maintain. it is "makeshift"
ent1ty wrote: success is a matter of concentration and desire
at a cost measure in computer resourcesButler Lampson wrote: all problems in Computer Science can be solved by another level of indirection
Yes yes, but maaan, two cubes collision! That's not a big deal is it?
Dreaming of folk-art at http://www.ambersky.cz
-
- Posts: 363
- Joined: Thu Dec 16, 2010 8:50 pm
- Location: Somewhere in the clouds.. drinking pink lemonade and sunshine..
no no no im not telling you or him to do anhything im just warning of having a bit of a rag-tag system set up. say he wants to expand whatever his project is, in future. and he needs to check collision between 100 cubes?andreee wrote:Yes yes, but maaan, two cubes collision! That's not a big deal is it?
he will have to tear his whole already established system apart to iuntegrate bullet or ODE or whatever collision lib you want anyways. so why not have a proper system for that to begin with?
well only the author knows what they really need. so im not one to speak..
ent1ty wrote: success is a matter of concentration and desire
at a cost measure in computer resourcesButler Lampson wrote: all problems in Computer Science can be solved by another level of indirection
I am sure that i won't need anything more that detecting cubes in my project (i think ).
Do you have a quick aid on how to detect those two?
EDIT: ok, after a bit of rethinking, how would checking collisions between say 100x100 cubes be more system-loading than checking that with a physics engine?
Do you have a quick aid on how to detect those two?
EDIT: ok, after a bit of rethinking, how would checking collisions between say 100x100 cubes be more system-loading than checking that with a physics engine?
Dreaming of folk-art at http://www.ambersky.cz
I'd say that irrlicht's collisions should not be used for that kind of stuff, selecting nodes from scene with ray is ok, but continuous collisions with many objects should be avoided.andreee wrote: EDIT: ok, after a bit of rethinking, how would checking collisions between say 100x100 cubes be more system-loading than checking that with a physics engine?
Working on game: Marrbles (Currently stopped).