Hemisphere Games Forum

Some levels start with everything on the field at a dead stop. Nothing is moving the slightest bit. Yet at the end of the level, you can encompass >99% of the level's mass and be moving at high speed!

What's happening is that you can get free acceleration from the immovable walls. For example, say you're travelling at speed and you want to come to a stop without losing any of your mass. Fire some mass out at the wall in front of you. This will slow you down. Then when it comes back at you, it will hit you head on and slow you down even more. This also works in reverse, putting no limit on how much you can accelerate (well, I'm not sure about "surface tension" effects of balls colliding), even given a small amount of reaction mass.

The trick will still work in the real world - think of an astronaut with a basketball outside of the space station. If he wants to zoom off into space, he can bounce the basketball off the space station over and over. Throwing the ball accelerates him backwards, and so does catching it. But the reason he's accelerating backwards is because the very massive space station is experiencing an equal force and accelerating a tiny bit the other direction. So the acceleration isn't free.

So do the same thing for Osmos. When mass hits a wall, accelerate the walls! This doesn't have to be difficult- you can equivalently keep the camera locked to the reference frame of the walls and accelerate all the objects the opposite direction. The mass term cancels out so you don't have to pick an arbitrary value.

The end result is that while it would probably take an exceptionally sharp eye to notice the tiny acceleration, when you're done with the level and hold 100% of the mass, you will be at a dead stop at the level's center of mass. I think this would be really cool; no matter how chaotic and crazy you play, and no matter what angle and pattern you fire mass, each collision cancels out a certain amount of its energy and the level reduces to a single coordinate and mass.

Do you mean, if you bump off the top wall, the entire box starts to move upwards, relative to the stationary motes?
Interesting Idea,I can see how it would unbalance things though. as it moves up, the motes near the bottom will bunch up, and with two directions of movement, or a circular board, enough movement would easily merge an ubermote. this could easily be done by a determined AI, or just a careless one.

Well, some physical laws had to be broken to create gameplay value.

yes!!!!!! reference frames! that would be a kooky game if the background had a variable velocity. I'm sure there are difficulties associated with this though.


I had a similar idea for reference frames when thinking about the epicyclic orbit levels. If there was some command like alt+f to switch between frames it would be neat. Most levels that I've played are in the frame of the...box. haha, but it would be really messed up if you could see things in the frame of the protagonist mote, or a specific attractor which you are orbiting for certain levels. For example, the orbit levels (which I've played so far) are in the reference frame of the central attractor. I would like to play the game with reference options.

fred104 wrote:yes!!!!!! reference frames! that would be a kooky game if the background had a variable velocity. I'm sure there are difficulties associated with this though.


I had a similar idea for reference frames when thinking about the epicyclic orbit levels. If there was some command like alt+f to switch between frames it would be neat. Most levels that I've played are in the frame of the...box. haha, but it would be really messed up if you could see things in the frame of the protagonist mote, or a specific attractor which you are orbiting for certain levels. For example, the orbit levels (which I've played so far) are in the reference frame of the central attractor. I would like to play the game with reference options.

You would love http://planets.homedns.org/
It's not a game, just a planet physics sandbox. You can add planets of different sizes and they'll interact realistically. All sorts of cool stuff works like stable Lagrange points. And the coolest part is that you can lock the camera over the center of mass of any group of planets, or just one. So locking onto a satellite makes it look like the planet is orbiting it, and locking onto a binary system gives you a neat little shuffle

never have I been so glad to have recently set up cygwin XD

Yeah, Catinthewall is right: if the player was able to impart momentum to the level boundaries, they'd move and gradually collect and sweep the level's motes against the far wall. These motes would tend to encounter one another and absorb each other, resulting in a small number of larger motes that would probably be too large and tricky for the player to deal with.

Interesting idea though! Instead of being able to move the outer walls, it might be interesting if there were free-floating wall pieces in the center of the level that the player/AI motes could shove around, impart momentum to by bouncing off, etc. Crazy!

Thanks for the link, bmn -- mmmmm, physics is so much fun!

Dave

One small note, you cannot gain infinite speed by bouncing off the walls. This is limited by the relative speed of matter expelled from your mote.

renoor wrote:One small note, you cannot gain infinite speed by bouncing off the walls. This is limited by the relative speed of matter expelled from your mote.

I'm not certain, but I think you're wrong. Imagine this scenario: you expel matter behind you (gaining speed), and your ejected matter bounces off the wall at a slight angle. You coast, going straight, until you hit the far wall and bounce off it. On the way back, you pass your ejecta without absorbing it. You bounce off the first wall, and (with a tiny adjustment in direction), come up behind your ejected mass and scoop it up. Now you've gotten the benefit of the full impulse going forward, but you only have to "repay" part (half?) of that impulse to get the mass back, since it's moving with you.

I doubt this particular method is practically feasible, but it shows that this is possible in theory. Maybe if you performed the same maneuver on each side of the area so that the angle-shots cancel out? Or maybe you can repeat it on one side, leaving your course uncorrected and not scooping up that ejected mass, until you hit one of the perpendicular walls, which corrects your direction for you. I don't know