Monday, June 18, 2007

Quiddle Me Vis #2

Although I haven't yet exhausted the recommendations from my last appeal, I am running a bit short on short things to do. With that in mind, I'm bringing this up again. Here's how it works this time: Ask me any question. If it's a silly question, you'll likely get a silly answer. If it's a serious question, you'll get a serious answer.

Whether you're puzzled over the definition of a quasar or you want my recommendation of which Terry Pratchett book to read first, now's your chance to ask. So get to it!

19 comments:

Theo said...

What's your view on free will? Do you think you have any choice in what your view is? Maybe you hold both views (or the potential for both, or perhaps a 60/40 split), which will split the universe the moment you decide to tell us!

And which Pratchett do you recommend reading sixth (I think)?

Bronze Dog said...

Why do you have to steal my act?

Anonymous said...

I'm terminally ill with Thripshaw's disease. Which Terry Pratchett book should I read last?

Infophile said...

What's your view on free will? Do you think you have any choice in what your view is? Maybe you hold both views (or the potential for both, or perhaps a 60/40 split), which will split the universe the moment you decide to tell us!

Ah, the mingling of free will with quantum uncertainty. Now, that one deserves a post of it's own.

And which Pratchett do you recommend reading sixth (I think)?

I'd have to know which ones you've already read, but a few general recommendations:

Pick a sample from each of his major story arcs (check out Wikipedia's list to see how they break down) and see which one you like most. Then just read those in release order. General fan opinion is that the City Watch/Vimes arc is the best, while the Wizards arc is the weakest. Outside of the arcs, a few stand-alone stories are quite good too. That I've read, The Truth and Monstrous Regiment were quite good, and I've also heard good things about Pyramids.

Why do you have to steal my act?

Believe it or not, I was trying for something a little different here. The goal was to get more serious questions that people were wondering about and that I might be able to extend into some post (and to challenge my intellect if I get anything really juicy). I did, however, choose to add in a note about silly questions as a tribute to your posts of that nature.

I'm terminally ill with Thripshaw's disease. Which Terry Pratchett book should I read last?

Reaper Man seems morbidly appropriate.

Bronze Dog said...

Believe it or not, I was trying for something a little different here. The goal was to get more serious questions that people were wondering about and that I might be able to extend into some post (and to challenge my intellect if I get anything really juicy). I did, however, choose to add in a note about silly questions as a tribute to your posts of that nature.

Why can't I get a silly answer? ;)

Infophile said...

Because I know you're secretly K-9 and you won't admit to it!

Nochte Elphi said...

Could you consider it possible, given unlimited resources and time, to create a computer which could perfectly simulate every physical aspect of the universe without being as large as the universe?

In other words, given the capabilities, do you think man could re-create this universe more efficiently than it has arisen on its own?

Infophile said...

Could you consider it possible, given unlimited resources and time, to create a computer which could perfectly simulate every physical aspect of the universe without being as large as the universe?

Now, this is one I could really dig into. But for now, a quick answer: Yes, but only if you cheat. That is, you could set up an encoding system which allows you to represent a complicated system (the universe) by a less complicated set of data (the computer). This allows you to represent a specific universe from a computer with the trade-off of not being able to represent any arbitrary universe. Additionally, you'd have to make the encoding system either ridiculously complex itself (rivaling the size of the universe) or be very lucky.

Nochte Elphi said...

Well of course if you could encode the universe it would be smaller, but do you think you physically could cheat?

Infophile said...

If you're under no obligation to actually built the encoding system, then yes. For instance, I'll represent this universe (past present, and future) by the number 1 with a hypothetical encoding system that will translate it into every aspect of the universe. Of course, coding like this is utterly useless.

Now, one way it might actually work is if we somehow get lucky with the state of the universe. For instance, if the universe were finite in size and completely empty, we could represent this with a computer that simply has a few parameters stating the size of the universe and that there's nothing in it (let's pretend for a moment the computer is outside the universe). This also works if the universe is periodic or happens to have some pattern that can be expressed mathematically by some equation. We'd only need to represent a subset of it with the computer and then explain its periodicity in that case.

However, the problem with those cases is again that while they can represent specific universes, they can't be used to represent arbitrary universes. Let's talk about ways we could do that.

The first problem we run into is what happens if the universe is infinite. If this is the case, and there's no periodicity or equation that allows us to cheat, then we'd need an infinite amount of information to simulate it. In order to store an infinite amount of data, we'd need an infinite amount of space, so the computer ends up being just as infinite as the universe. The end result here is that if the universe is infinite, it's flat-out impossible to simulate it with a finite computer.

So, let's stick to finite universes for the time being (and let's also say the computer is outside the universe in order to avoid infinite loops of explanation). One solution that immediately comes to mind is making a scale model of the universe. The problem here is that it would require altering a few physical laws so everything would work out the same way but on a smaller scale. If we're allowed to do that, we have a solution. If not, we'll have to try to get creative.

The next obvious step is to look at the universe as akin to some data file, and the computer as a means of compressing it. The problem we run into here is that many data files simply aren't compressible - they're too complex to start with. The cases where compression is possible are akin to the simple cases discussed previously. In the end, we can compress some universes, but not all.

But let's say for the moment that we can compress data of the universe. If the universe were completely deterministic, a snapshot of it (or rather, a movie of it for a very short length of time) would be sufficient to completely construct the past and future of it. However, it seems that Quantum effects introduce an inherent randomness to the universe, where many events end up being determined randomly as the universe progresses. For the computer to simulate the future of the universe, it would have to know the outcome of every one of these events. If the future is finite, we might just be able to get away with it.

However, in our universe, it's looking like the future is infinite. The universe is expanding and the expansion is accelerating. From our current understanding, it will go on forever. This leaves us with having an infinite amount of data we'd have to store in the computer, and to even get this data we'd have to travel in time.

Theo said...

Doesn't it depend on whether the universe is completely deterministic or not? Even if it is, and all the laws are fully understood and programmed (i.e. the algorithm for propagating the universe over time), wouldn't one still need to be able to "store" a complete representation of the universe at a given frozen moment in time?

On the other hand, if there is randomness in the "model" (maybe at the quantum level only?), one could still conceivably model and simulate this, but it won't be as useful for predictions...

Also, one would need to be able to store multiple paths over multiple moments (universe "snapshots") in time... Biiiiig hard disk required!

Nochte Elphi said...

Compressability is essentially what I was getting down to. If the universe we live in, I only use this one because it is the one we know most about, is compressable then this means that some areas of the universe are redundant. Since we would be dealing with the most fundamental aspects of reality, redundancy would mean that there would be multiple instances of the same thing. What I mean to say, is this would lead to having the same exact 'object' existing in two seperate locations in the universe.

I was under the impression that this is not so, atleast with our universe.

When I say 'object' I am reffering to actual matter as well as the effect it has on its surroundings and vice versa. If one tried to have two particles who were identical in every way, let us say particle A and B) A would be closer to some other hypothetical particle than B would be. This would correspondngly affect A differently than B thus differentiating the two.

Allow me to rephrase my original question. Do you think the universe we currently live in is compressable?

As for having the computer predict the future, I believe it could if it had information about the entire universe. However, it would by definition have to behave according to the same laws meaning that the computation of predicting the future would take as long as the "real universe" would take to get to the future.

Infophile said...

On the issue of compressibility, I'd have to go with "most likely no." The issue here is what's colloquially known as "Quantum soup," or the chaotic quantum events that go on everywhere, even in a vacuum. These events do have measurable effects, so we can't ignore them, and they permeate the entire universe at a nanoscopic scale. Given all this essentially random data, there's no way we'd be able to compress it without losing information.

Quantum mechanics also throws a wrench into determinism, unfortunately. Its results are innately random, so there's no way we could use a current model to perfectly predict the future.

However, in the hypothetical scenario that we had a deterministic universe, it actually is possible for us to have a universe which is periodic on large scales. This means that we could have identical objects A and B in different unit cells, and while they're different distances from object C, the corresponding objects to C in other unit cells when added together produce the same effects on A and B. However, there's no indication that our universe is like this, and it would seem extremely unlikely.

Nochte Elphi said...

the corresponding objects to C in other unit cells when added together produce the same effects on A and B.

How?

Infophile said...

Let's slightly modify the naming of these objects to make it clear. In the first unit cell, we have objects A and B. In the second unit cell, to the "left," we have objects A' and B', which are identical to A and B.

Now, A and A' are different distances from B, but the distance from A' to B' is the same as from A to B. And then the effect of B' on A would be the same as the effect on A' from some B'' in the next unit cell down. And then to match the effect of B'' on A, there's a B''' for A' even farther away. If the unit cells stretch to infinity or wrap around to connect to the beginning, all of these effects will add up to make A and A' affected exactly the same way.

Nochte Elphi said...

I see. Well taken, but I shall remain a stickler and note that this obviously is only applicable where this string or circle of particles is the only thing in the particular universe. All you need is to have some particle outside of this string or ring to distort the ballet of forces.

However, I give you props on your hypothetical situation

Infophile said...

Oh, of course this is extremely unlikely to be the case. In fact, in observing such a universe, it would be indistinguishable from one which was simply self-contained (the fabric of space wrapped back to connect with itself). If you were ever in a position to be able to tell the difference, you'd also be in a position to alter different unit cells in different ways, and this would then break the periodicity.

Anonymous said...

Which Terry Pratchett book should I read 12th?

Infophile said...

Take whichever book you read eleventh, and read it again with the chapter order reversed. The jokes stand out more this way.