what is a computer? how does it work?

[Michali]

What essentially is a computer? Like if you had only two sentences to describe it, what would they be? I was thinking:

A computer is tool by which electricty is discharged in a satisfying manner; a manner which is dictated by the state of the tool itself. The state of the computer is a result of the manner in which the discharge occurs, and this discharge can occur in many more ways than one.

The other question is how does a computer essentially work? I don't know at all how it works. All I'm guessing is that it is a tool which taps into a power source, and completes circuits in order to record binary changes to a material. Then those changes are used in summation in order to create binary changes to a new material ( a new device ), and this process continues until changes are made in the intended device (like a monitor).

So, if my task is: display a white equilateral triangle on a monitor, what exactly happens to make that occur?

[Wonderbread Leotard]

Originally Posted by Michali

What essentially is a computer?

A device that takes in, processes, and outputs information. In the broadest sense computers need not be electrically powered and even existed in the ancient world.

[naturalist.atheist]

Originally Posted by Michali

What essentially is a computer? Like if you had only two sentences to describe it, what would they be? I was thinking:

A computer is tool by which electricty is discharged in a satisfying manner; a manner which is dictated by the state of the tool itself. The state of the computer is a result of the manner in which the discharge occurs, and this discharge can occur in many more ways than one.

The other question is how does a computer essentially work? I don't know at all how it works. All I'm guessing is that it is a tool which taps into a power source, and completes circuits in order to record binary changes to a material. Then those changes are used in summation in order to create binary changes to a new material ( a new device ), and this process continues until changes are made in the intended device (like a monitor).

So, if my task is: display a white equilateral triangle on a monitor, what exactly happens to make that occur?

There was a time when a computer was a person that sat in a large room with other people called computers and what they did was perform computations like addition, subtraction, division, multiplication, square roots, logarithm, sine, cosine and so forth. Early on in the Manhattan project those kinds of computers were used to design the first atomic bomb. They were managed by a young physicist named Richard Feynman. At around that time the first electronic computers were developed in England and the US and were also used to perform computations for the A-Bomb.

In order to compute you must represent the thing you want to compute in numbers. Then you must perform operations on those numbers. A particular computation will take a particular recipe of additions, subtractions and so forth performed in a particular order using information stored as numbers and taking intermediate results and storing those as numbers in some sort of a short and long term storage location. This recipe is called a program.

Now the internals of a computer are rather interesting because numbers are just representations. There are many ways to write or express a number. The computer does it with electrons and the operations and programs are accomplished via electronic circuits. A circuit being a collection and configuration of electronic devices such as transistors, resistors, diodes, capacitors, and so on to accomplish a specific task. The basic circuit in a computer is the switch. Not too different in function from the switch on your wall used to turn your lights on and off. Except these switches are controlled by other switches and what is switched on and off is usually a rather minute voltage or current. In modern microprocessors in use today the voltage is used to represent the most basic of counting unit, the 1 and the 0. The 1 being when the switch is on and the zero being when the switch is off. By taking a collection of switches you can represent a number. So three switches can represent a number from 0 to 7 as 000, 001, 010, 011, 100, 101, 110, and 111. And you can create a circuit that would take two collections of three switches and add them together and place the result in a third collection of three switches. In other words a circuit that would take 010 stored in location A and 001 stored in location B and place 011 in location C. A computer is a collection of billions of such circuits all hooked up together in this amazing coordinated machine that operates in precise lockstep cycling at billions of times a second with close to absolute perfection.

This creates a device that is capable of performing astounding feats of computation, allowing you to represent all sorts of things like games, buildings, or even the universe and compute its behavior in great detail as a series of arithmetic operations.

[Dragar]

Originally Posted by naturalist.atheist

There was a time when a computer was a person that sat in a large room with other people called computers and what they did was perform computations like addition, subtraction, division, multiplication, square roots, logarithm, sine, cosine and so forth.

Entirely correct, but it is worth mentioning that these people were predominantly women, and the scientists predominantly men.

[naturalist.atheist]

Originally Posted by Dragar

Entirely correct, but it is worth mentioning that these people were predominantly women, and the scientists predominantly men.

It was a very different world back then. I am very glad for my daughter that we do not live in that world. But for a large portion of the population of the world today that old world still exists. I commiserate with their daughters. It is a huge waste.

[Dingfod]

A Babbage computer:

[naturalist.atheist]

You can also build logic circuits with fluids like water. In 1964 the first all-fluidic computer was built by UNIVAC.

[Michali]

hmmm, I think what I mean to ask is, putting terms like "information" and "numbers" aside, can we describe what is happening in purely physical terms? Like, for any thing considered to be a computer, what is happening physically?

-edit- also, thank you naturalist atheist

[ceptimus]

A computer models information and numbers according to how it is designed, so it's not meaningful to say what is happening in purely physical terms without refering to what the designers decided that those physical changes mean.

For example, consider a computer designed to (amongst other things) add numbers, and numbers are represented by the height of water in glass tubes. So we might have three inches of water in one tube and four inches in another and to add the two together the computer would open valves at the bottom of each tube and let the water run into a third empty tube of the same diameter. The water in the third tube would be seven inches high representing the answer 3 + 4 = 7

Without an understanding that the height of water in a column represented a number, any explanation of valves opening and water flowing from one tube to another would be fairly meaningless.

[Wonderbread Leotard]

Originally Posted by ceptimus

A computer models information and numbers according to how it is designed, so it's not meaningful to say what is happening in purely physical terms without refering to what the designers decided that those physical changes mean.

True. And the definition I posted is quite vague. But it beats out a definition of computer I saw in a dictionary -- "a device that computes." Gee, thank you dictionary, I never thought of that!

[seebs]

A computer is a device which translates inputs into outputs according to a particular set of rules. The rules can be very simple, or very complicated. There exists a small set of basic operations which, if you can perform them, you can do anything we know of that any kind of computing device can do. This was established by Alan Turing, and a device which can do this is called "Turing complete". In most cases, a programming language is viewed essentially as a virtual device which provides a particular set of rules for operations, and thus one speaks of a language as being "Turing complete" or not. (Nearly all are.)

[Michali]

Originally Posted by seebs

A computer is a device which translates inputs into outputs according to a particular set of rules. The rules can be very simple, or very complicated. There exists a small set of basic operations which, if you can perform them, you can do anything we know of that any kind of computing device can do. This was established by Alan Turing, and a device which can do this is called "Turing complete". In most cases, a programming language is viewed essentially as a virtual device which provides a particular set of rules for operations, and thus one speaks of a language as being "Turing complete" or not. (Nearly all are.)

Are there any common physical properties to Turing machines? Or are you restricted to use of terminology like "rules" "operations" and "translate"?

[seebs]

It's purely abstract terminology, there are no common physical properties, not even the requirement that there be any physical properties. They can be electronic, they can be mechanical, they can be purely conceptual, they can be biological, doesn't matter.

[Michali]

Originally Posted by ceptimus

A computer models information and numbers according to how it is designed, so it's not meaningful to say what is happening in purely physical terms without refering to what the designers decided that those physical changes mean.

For example, consider a computer designed to (amongst other things) add numbers, and numbers are represented by the height of water in glass tubes. So we might have three inches of water in one tube and four inches in another and to add the two together the computer would open valves at the bottom of each tube and let the water run into a third empty tube of the same diameter. The water in the third tube would be seven inches high representing the answer 3 + 4 = 7

Without an understanding that the height of water in a column represented a number, any explanation of valves opening and water flowing from one tube to another would be fairly meaningless.

Right, so this is very well put and interesting. I'm also inclined to call this a computer. In fact, it is a calculator. We can just put one number in tube A, and another number in tube B. Open the valves and you get a total number in tube C. So the entire device is a computer.

So a computer is just a tool. No doubt, some outputs of computers turn on lights, and unlock doors. But is there no commonality to all of these differents types? Or is it that they -could- do different tasks?

Let me try to define it further by associating its simple "toolness" to something else. A computer is a simplifier. The tool takes physical aspects that we don't well-know, and it makes them into phsyical aspects that we do well know.

A computer is a tool which turns complexity into simplicity.

An electrical door-lock, no doubt goes from a complex set of interactions, to one action which is unlocking the door. The liquid calculator is the same way. Also, impressions from a hard drives, are extracted from their locations on the disc, and impressioned on the memory card. And from there, to the video card.

So, once again, a computer is an object which converts complexity to simplicty.

[naturalist.atheist]

Originally Posted by Michali

hmmm, I think what I mean to ask is, putting terms like "information" and "numbers" aside, can we describe what is happening in purely physical terms? Like, for any thing considered to be a computer, what is happening physically? -edit- also, thank you naturalist atheist

What is happening is not so much physical but representational. However the representation takes on some physical form but once you decide what that is, it almost becomes irrelevant to the computer. (In theory of course but the devil is always in the details and usually a great deal of work is done to implement the core processing elements needed for any digital computer be it a water computer or an electronic computer or an electron spin computer.

Take the fluidic computer. It works by using the flow of a liquid in tubes with various fluid control devices that work very much like logical gates that can manipulate zeros and ones. In that computer a 1 could be the presence of a flow of the fluid and a zero the absence. Here is a link to an enterprising student that used the flow of water to create a simple computation device that operates on a single bit of information. Once he decided how the flow of water would be used to represent information he then figured out how to construct fluid flow devices that would process that information as a logical AND and XOR.

[seebs]

Nope. It's what we said it was, not something else you think might be similar in some way. There is no "converts complexity to simplicity" going on in general -- there's some cases that look like that, and some that look the other way. (For instance, look closely at "1 + 3 = 4", and then look at the many, many, thousands of gate flips which are involved when a modern computer does that calculation.)

It's just a device which transforms inputs into outputs following rules. Nothing more, nothing less.

[Dingfod]

A computer is a device that connects you to the bundles of tubes known as the intarweb or something like that, where people can hack your IP and make phone calls in your name, or something like that.

[naturalist.atheist]

Originally Posted by Michali

hmmm, I think what I mean to ask is, putting terms like "information" and "numbers" aside, can we describe what is happening in purely physical terms? Like, for any thing considered to be a computer, what is happening physically?

-edit- also, thank you naturalist atheist

In looking at this question again, there is a kinda yes answer. Before the advent and ubiquity of digital computers there was a kind of computer called an electronic analog computer. It used voltages and currents to represent data and results however it did not process the information in anything resembling digital computation. It consisted of a series of operational amplifiers that could perform complex operations like differentiation, integration as well as the simpler arithmetic operations. And the speed of processing was limited by the signal settling times of the operational amplifiers used. Because the data was represented as a voltage a value of 5,879 might be represented as .5879 volts. If the answer was 16,780 then the final output would read 1.678 volts. What you would do is wire up these operational amplifiers with resistors, capacitors, inductors and so forth into circuits that would perform the basic operation. If the calculation was a series of operations then you would create a circuit for each operation and chain the output of one stage to the input of another. You would then apply the input voltage to the input of the first stage and then read the output voltage of the last stage. If the settle times of the Op amps used was in the micro-second range and you had staged 10 Op amps then about 10 micro-seconds later the voltage of the output stage would settle to a more of less constant value and that would be your answer. The explanation of the computational process was done purely in physical terms in that rules of analog circuit design was how you would explain the computation.

That type of computation is still widely used today but not for computations of any complexity or accuracy since each stage could have a fairly large error introduced by tolerance variations in the various components used to build the stage. Whereas a purely digital machine does have a round off error but other than that it is pretty much perfect.

[naturalist.atheist]

Another example of a form of "computer" that would use a mechanism of computation that could be described in purely physical terms would be an example of a crude multiplier based on a simple fulcrum. If you wanted to multiply 2 x 4 you could arrange a beam with a fulcrum placed at a mark on the beam labeled 4 (this mark would be at 1/5th the length of the beam). You would place two 1 pound weights on the long end of the beam, you would then go to the short end of the beam and start placing 1 pound weights until the beam balanced. Count the number of 1 pound weights and that is your answer.

Another example would be an adder built with two straight rulers. Say you wanted to add 5 and 3. Take one ruler and place the starting edge at 3. Look along that ruler until you come to five. Look across from that ruler to the other ruler and see what number lines up. It will be 8.

Another example which my daddy taught me when I was a small child was a simple multiplication machine using a piece of paper and an "L" cut from cardboard or paper. On the first piece of paper draw a 10 by 10 grid of boxes. (you can extend it as far as you have room). In the top row of the grid start on the left and put a 1 in the left topmost grid box. Put a 2 in the box to the right on that row and continue to the next box with a 3, 4 and so on. Go to the leftmost row and in the box below the 1 put a 2 and proceed down to the next row with a 3, 4 and so on. The next stage you can do by rows or by columns. On the 2 row and two column place a 4, then for the next column add 2 to the last entry and place a 6, add 2 again and place an 8 and so forth and so forth. For the 3 row and two column place a 6, add another 3 for the 3 column and place a 6, and so forth and so on. When you are done you should end up with something that looks like this:
1 2 3 4 5 6 7 8 9 10
2 4 6 8 10 12 14 16 18 20
3 6 9 12 15 18 21 24 27 30
4 8 .... fill this in
5 .... fill this in.
... fill this in.

Then take a piece of paper an inch larger than the size of the grid and cut out an L going along the outer edge. Then to multiply 8 x 3 you line up one edge of the L at the 8 on the top row and the 3 on the left column and the number in the corner of the L is the answer, 24. You have another kind of multiplication machine. I leave it to the reader to figure out how to use this as a divider.

[Michali]

The cardboard computer was also interesting. You said that it is less physical, more representational. Would the game of chess be considered a computer game under this heading? I would think so.

But I actually think that it is less representational and more perceptual, if I could draw a distinction. For instance, if the game of chess could be considered a computer game, then you would have to include the operations in your brain as being part of the computer. All forms of representation will include operations of the mind. From the shape of an equilateral triangle on a monitor to a number.

So all computers, at least of the calculator sort, will physically end within the processes of the user's brain. Other computers, of the non calculator sort, will unlock car doors, etc. I have to cut this short, but are those non-calculator computers using representational objects?

[Michali]

Originally Posted by seebs

It's just a device which transforms inputs into outputs following rules. Nothing more, nothing less.

So, would you say that having a "computer" is the same thing as having a set of "rules"?

[ceptimus]

A computer does rigidly follow a set of rules when it is operating, but it's different to, say, the rules of chess. The rules of chess define what moves are legal, but they don't tell you exactly what move you must make.

If you had some way of defining exactly what move must be made, leaving no choice to the player, that would be more like the way a computer operates.

When you do something according to a rigid set of rules - such as long division - you are working like a computer.

[naturalist.atheist]

Originally Posted by Michali

The cardboard computer was also interesting. You said that it is less physical, more representational. Would the game of chess be considered a computer game under this heading? I would think so.

It is interesting that you make that observation. Johnny Von Neumann, who is well known for his many contributions to physics, biology, mathematics, computing and in particular game theory called chess a form of computation. He did not consider it a proper game like say poker.

The lever and ruler computer were less representational and more physical in that the way in which one explained how the computer works relies on well known geometric and physical explanations. The cardboard computer is not a computer per se but a retrieval device that happens to retrieve the answers for multiplying a given pair of numbers.

But I actually think that it is less representational and more perceptual, if I could draw a distinction. For instance, if the game of chess could be considered a computer game, then you would have to include the operations in your brain as being part of the computer. All forms of representation will include operations of the mind. From the shape of an equilateral triangle on a monitor to a number.

So all computers, at least of the calculator sort, will physically end within the processes of the user's brain. Other computers, of the non calculator sort, will unlock car doors, etc. I have to cut this short, but are those non-calculator computers using representational objects?

I'm not sure I follow you on your use of the term "representational objects".

[seebs]

Originally Posted by Michali

So, would you say that having a "computer" is the same thing as having a set of "rules"?

No.

[Michali]

Originally Posted by seebs

No.

So a computer follows rules? So the rules are completely separate from a computer?