Monday, December 05, 2005

Visualizing Quantum Dynamics

It's not often that I throw a book away as worse than useless.

But I did over the weekend. The book was "The Dancing Wu Li Masters", a book on quantum physics for "reasonably bright non-scientists". I read about seventy pages before I could not continue.

The science was good, but his commentary on scientific culture, the nature of science, and how quantum dynamics "applies" to "real life" were all hideous. As just one example, he mixes his quantum metaphores and comes up with "we make the universe by how we look at it".

An idea later popularized by the mockumentary "What the !@#$ do we know" ("!@#$" theirs, not mine), this mis-use of quantum theory, along with similar mis-uses, leads to psychics and self-delusional people believing it is "evidence" that "explains" "psychic" "powers". Gosh, I love quotes.

This way of thinking is a simple mistake that anyone can make - and many people do. When confronted by quantum mechanics, they try to relate it to something they've experienced. Like say, an apple or a basketball game. This leads to a hodgepodge of explanation which is half gibberish and half flat-out wrong. They accept the half-gibberish part because they've been told that quantum physics "can't be visualized". This means they also accept the wrong part. They start thinking that the universe doesn't exist when it isn't observed, can be made in specific ways as you start to observe it, and, if you stretch a bit, can allow for all sorts of neat telepathic and multiple-existance phenomina.

Well, the book is wrong, the movie is wrong, and those people are wrong.

It's a problem with visualization. Since they have never been taught a way to visualize it, they end up just making random assumptions that sound neat. After all, everyone knows quantum physics can't be visualized.

Well, I'll tell you how I visualize it. This may not be something you can use to define new theories of quantum dynamics, but it is something that can prevent you from making the more grotesque errors in scaling and theory.

You can't apply quantum physics to your daily experiences, they say. Sure you can. You just have to look at daily experiences that are a bit further from the norm.

Fortunately, here in the game industry, providing experiences a bit further from the norm is what we do best.

Imagine an RTS. Like, say, Starcraft. As the players play, they uncover more of the map.

Now imagine you are watching the match. But you can't see the players: all you can see is how much of the map they've uncovered. And you can only look once per game.

So, if you look at their uncovered map right at the beginning, you know pretty much exactly where they are. They're in that tiny dot of discovered space. You can't really tell what they're going to do next, but at least you know they're going to do it around there. If you wait until much of the game has gone by, you see that they have discovered much of the map. You can't tell where they are in the area they've discovered, but you can see how their battle strategy is developing and in what direction they are moving.

Of course, Starcraft is more like a nuclear explosion than a careful watching of one particle. After all, the player builds bases and soldiers from the resources in the ground. So, let's ditch the whole "base" idea, and say that each player is only one unit, and the playing field is extremely large and a bit maze-like.

Now when you look at the map, you see a kind of squished spider web. Tendrils run all around, connect up seemingly at random, shoot off, stop, turn, and otherwise make a jumble of discovered terrain.

Where in that map is the player? It's impossible to tell. Players constantly double back, scout around interesting locations, and give up on a particular path to try another direction. Maybe they're in one of the tendrils, but which one? And even if they are, there's no telling exactly where they will be next - they could turn and go off at an angle if there's a lake in the way.

But looked at as a whole, the map of the area they've discovered paints a picture of movement in a particular direction. Some players will wend their way out circularly, others will strike out in a particular direction and keep going that way, and still others will gladly glance off in any direction at anything vaguely interesting.

You can tell how fast the player is going in what direction, but you can't tell exactly where the player is... or where he will be.

On the other hand, if you look at the map after the player has been painting it for only a short while, you see more exactly where he is, and less exactly where he is going. Remember: you can only look at the map once in a long while. So you can't get it once per second and watch it grow: by the time you look again, the fast-moving single player will have changed the known map utterly.

That's the essence of the problem.

Now, let's say that two players meet. Their maps can overlap for an arbitrarily long period of time before they encounter each other, but in this case, let's say they do.

What happens? Depending on the players, they might move away from each other. Or they might barely notice one another. Or they might even hook up, if their attributes are compatible. Like, say, two hydrogens and an oxygen would hook up. Or a long protein strain.

As a platoon, they stay near each other. Protein folding is nothing more than the various units jockying for the best tactical position.

These platoons cut a much wider swath of map and don't move any faster. In fact, they move slower, as team members scout nearby and report back, slowing up the march of the whole platoon. If we look at them on the same scale we were using before, it no longer looks like tendrils, but like blobs and fat pseudopods. So we can scale back. Zoom out.

Hey, zoomed out, the path doesn't appear so tenuous. Sure, it's kind of shaky, but it's easy to see which direction it is going in. Also, because everything is moving slower, the changes in the map between your glances are becoming slightly more maneagable.

Now, lets say those molecules match up with each other and become a group of linked molecules contained by a perimeter. Functionally, an army. Or a bacteria. Whichever.

This new group cuts tremendously large swaths. These things paint a path so wide that it would encompass the whole screen we used to watch a single-player's wanderings. Also, the group dynamics are even more complex than with a multi-player molecule. That means that the path of the whole group is significantly slower than the path of the individual molecules. (Assuming outside reference, of course. Otherwise, motion is meaningless anyway.)

This group is moving slowly enough that the changes to map between our looks are relatively minor. We can clearly see, "oh, they've explored a little bit in that direction since last time. So we know that they are there, and we know where they are going."

As you can see, quantum mechanics does not scale because big things move slow and cut wide paths.

I'm not pretending this visualization is ideal, but it's the best I've heard, and it should show you why you cannot use quantum physics on a large scale. The math simply doesn't work out: the whole reason that you can see location or velocity, but not both, is because of the scale you're working on. Once you start turning loose subatomic particles into atoms, the scale gets larger, slower, and easier to see. Once you start turning atoms into proteins, even more so. Once you turn proteins into cells, even more so.

Of course, throughout this entire visual exercise, I've only covered one element of one theory of quantum physics. You can try to adapt this model to the observer theory, or photon emission, or whatever you like.

It's fun. :)

No comments: