Exploring the Connection Between Physics and Computer Science

[Chrono]

I'm a computer science major and I was talking to a senior and he was doing a program for a presentation and he was explaining it to me. He went on about third and fourth dimensional arrays and he said that I will understand it better when I take physics. I didn't realize that physics had a lot do to with computer science. Do these two relate a lot?

Jon

 

[chroot]

They don't relate much, really -- but there is an entire branch of physics known as computational physics. It deals with the simulation of very complex systems, like supernovae.

And of course, computers are used to analyze data and to solve equations. Physicists commonly use tools like Mathematica to analytically solve problems like Einstein's field equations. Physicists also use a lot of computer processing to reduce and statistically analyze their data. The large particle accelerators could never produce useful data without tremendous computing power.

- Warren

 

[Anttech]

In my opinon Computer Science is totally dependent on Physics. Physics was what brought Silicon Computer to our lifes...
All Network equiptment was developed though physics (IE Cat5 through the application of laws governed by Physics)...
A good Computer Scientist has to be able to think in same abstract way a Physists and mathematician think...
Quantum Physics is being implemented into PC arcitecture now with the use of Quantum logic Gates...

The list goes on and on...

 

[chroot]

Anttech,

That's not computer science, that's computer engineering -- two very disparate fields, two distinct degrees, and so on. Computer science is essentially a branch of mathematics, and studying algorithms, optimizations, and the like. It has nothing to do with physical reality. Computer engineering, on the other hand, is like all engineering: essentially applied physics.

- Warren

 

[Anttech]

Fair enough, I aggree that's a better definition.

I have seen the Label "Computer science" used for the Engineering aspect of Computers and the hardware accosiated with it... When talking about degrees there is a definete distintion...

I am by trade a Network Engineer, and for our technology to mature we definately need Physics. We also need Mathematics for maturing encyrption technics and refining how data is put more efficently onto the wire...

 

[BigRedDot]

Computer science is essentially a branch of mathematics, and studying algorithms, optimizations, and the like. It has nothing to do with physical reality.

The folks at Caltech's IQI, might disagree with this last sentence, I think. Although I am somewhat dubious myself that a useful quantum computer will ever come to be built, I recognize the distinct posiibility that even so, some of the central questions of computer science might be answered by investigations into quantum information theory. (In fact, it would tickle me pink, and maybe even purple, to see a P!=NP proof show up in Phys. Rev.) One of the things a quantum computer (should a working one ever be constructed) could be good at is efficiently simulating quantum systems. On the other hand, this really blurs the lines between "running a calculation for a model of an experiment" and "just observing the experiment." While I am not a mathematical platonist by any stretch, I am not so sure that information itself is so divorced from reality as mathematics is.

 

[slyboy]

> Computer science is essentially a branch of mathematics, and
> studying algorithms, optimizations, and the like. It has nothing to > do with physical reality.

Indeed, but it is a branch of mathematics that is supposed to relate to what real physical computing machines are like. There is a whole list of assumptions about what is a "physically reasonable" model of computing that have to be determined before one can even start studying algorithms or complexity.

As an example, one might ask why we don't take analog computers as our fundamental model of computing rather than digital computers. Analog computers are much more powerful in theory, but it would take an exponential amount of energy to utilize this power. They are therefore ruled out by an essentially physical argument.

The founders of computer science thought that the assumptions they were making were obviously correct and that they did not have much to do with physics. This was bourne out by the fact that every "reasonable" model of computing proposed until recently can be simulated using a Turing machine with at most a polynomial time overhead.

However, Quantum computing is not just another area of computer engineering. It has its own entirely different theory of algorithms and complexity, which are potentially much more powerful than the standard ones. Theorists (both physicists and computer scientists) have spent a lot of time trying to justify why this is a "reasonable" model of computing, but the ultimate test will be whether the "quantum computer enigneers" (who are actually mostly physicists) can manage to build one of the things.

On another point, there are some physicists who think the connection goes the other way round as well. Indeed, Wheeler coined the phrase "it from bit" to express the idea that some modern physics theories are basically a kind of information theory. Black hole thermodynamics and quantum physics are the two main areas to which these ideas have been applied.

 

[MisterBig]

Going back to arrays...

An understanding of physics or maths is not a prerequisite for imagining a 4d array.

A 1d array can be imagined as a line of boxes (each box containing some piece of information). Reference by box number
A 2d array is a stack of these information boxes. Reference by number of boxes from top and number of boxes from left
A 3d array sees all these information boxes arranged in cube. Reference by number of boxes from top, number of boxes from left, and number of boxes deep

To imagine a 4d array all you need to do is put a line of boxes inside each box in the cube - Reference by number of boxes from top, number of boxes from left, number of boxes deep, and number of the box inside the box we just referenced.

Imagining a 4d array in 4 dimensions isn't really necessary. Having said all that, you will find that some knowledge of physics and certainly of maths will come in handy on your course.

 

[Chrono]

I'm guessing that physics will, overall, help with my understanding of computer science.

 

[StarkyDee]

I would like to see more replies on this thread because it is an important one for the future. I'm currently pursuing a computer science major with a biology/nanotechnology minor. This field of science is the most interesting to me i.e., molecular nanotechnology,biological mimicry, biotech, etc. Physics has so much to do with computer science it is unbelievable. Whether talking about the philosophy or the integration of engineering computers, physics plays a key role in development of smaller and smaller computers-hence having more information. I would like to know more about the Mathematica program and it's characteristics/applications/processes that it can handle. Any good sites on this program?

 

[chroot]

Originally posted by StarkyDee
Physics has so much to do with computer science it is unbelievable.

Not really -- only the fringes. CS is a branch of mathematics, and has as much to do with physics as does number theory.

I would like to know more about the Mathematica program and it's characteristics/applications/processes that it can handle. Any good sites on this program?

How about www.wolfram.com?

- Warren

 

[Chrono]

Originally posted by StarkyDee
I would like to know more about the Mathematica program and it's characteristics/applications/processes that it can handle. Any good sites on this program?

We use Mathematica at school for our calculus courses. As for sites, the only one I could recommend is the offical site of the program.

 

[BigRedDot]

Not really -- only the fringes. CS is a branch of mathematics, and has as much to do with physics as does number theory.

Whether computer science has much to say about physics is up in the air (although, some people -- Wolfram for example, but others as well -- would certainly say that it does). However, it seems evident that physics does have things to say directly about computer science, and not just the engineering aspects. Again, whether a functioning quantum computer ever comes to be built or not, the investigations into their theoretical capabilities impinges directly on computational complexity theory.

 

[Anttech]

Heres some info on controlled quantum logic gates for computers
http://www.infoworld.com/article/03/10/29/HNquantambreakthrough_1.html

 

[slyboy]

Not really -- only the fringes. CS is a branch of mathematics, and has as much to do with physics as does number theory.

You are obviously unaware of the strong connections between the Riemann hypothesis (possibly the most famous unsolved problem in number theory) and Quantum Chaos theory (a fundamental area of physics). In fact, many results about the Riemann zeta function have been derived using an approach based on quantum physics.

To play devil's advocate, there are really connections between EVERY area of human understanding. The borders that we errect between subjects are to some extent arbitrary and have been set up by us to enable us to learn and think about the world efficiently. The question should not be whether there are connections between different areas, but how fundamental those connections are to each of the subjects involved.

In the case of computer science, I would say that the connections are pretty fundamental. For example, look at the definition in complexity theory of what is supposed to be a "tractable" problem for a computer to solve. It is usually given by whether an algorithm exists to solve the problem with a number of time-steps that is polynomial in the size of the input. However, this definition is only really justified by Moore's law, which essentially says that the speed of computers increases exponentially. This guarantees that size of computation that you can achieve on a real computer in a reasonable amount of time increases appropriately if you have a polynomial algorithm. However, Moore's law is limited by physics, i.e. by how small components on microchips can possibly be, and many people think that it will not continue to hold for ever. If it starts to break down, then computational complexity theory will still be mathematically correct, but it will no longer be relevant for actual real physical computers.

On a more practical level, the big trend in scientific research, both academic and industrial, is to go for projects that are interdisciplinary in nature. Therefore, if you are going for this sort of career then it is a big advantage nowadays to have a grasp of more than one subject area. It is also much more fun to do this sort of work than to be one of those people who only talk to the three other people in the world who understand their subject area and are totally blind to any wider implications that their work might have.

 

[chroot]

Originally posted by slyboy
To play devil's advocate, there are really connections between EVERY area of human understanding.

Very true. I'll retract my previous statements a bit, since I seem to be the odd man out here. The original poster was mostly asking about CS & physics as taught in a university -- and you can bet very few of the fundamental connections you mention are explored at the undergraduate level in either program. As far as undergraduate-level academia is concerned, CS and physics are essentially completely disparate studies.

- Warren

 

[Chrono]

Originally posted by chroot
and you can bet very few of the fundamental connections you mention are explored at the undergraduate level in either program. As far as undergraduate-level academia is concerned, CS and physics are essentially completely disparate studies.

- Warren

You're probably right about that. We're not required to take physics, just a two semester science.

 

[JAugusto]

Computer sciences are completely based on physical laws and thus lack the "soul" aspect of the “complete reality”. That's what I think will be the biggest problem with making computers gain conscience one day. We merely understand a little bit of the “complete reality”, and even that little bit is very “flatlanded” distorted…


http://www.zapfuture.com


.

 

[Achy47]

Both fields have their own usefulness for the quantity of entropy.hehe. Thermal vs Logical. Thats about all of the similarity.. Just joking.

My personal opinion is that everything is related.

btw. If loop quantum gravity is experimentally verified, does that make EVERYTHING a discrete system?