English Deutsch Français Italiano Español Português 繁體中文 Bahasa Indonesia Tiếng Việt ภาษาไทย
All categories

Computers (as I understand them) operate based on an off/on behavior, aka base 2, as does most other "digital" equipment. The reason for this limitation is because a component (transistor) can be either on, or off. Same for fiber optics - there is either a pulse of light, or a lack of light.

Wouldn't things be faster and more advanced if you could operate things on a base 5, 10, 20, or higher?
For example, instead of fiber optics sending pulses of light, have a type of variable wavelength laser, which could produce 8-256 or more different colors based on the laser's wavelength. Of course they would have to develop the technology to both send and receive the beams... but it could be done. Imagine how much faster electronics with microprocessors would be!

The question is, why hasn't it been done yet (or has it?) What are the current drawbacks (besides the lack of hardware)?

2007-06-14 07:07:16 · 11 answers · asked by stevepg82 2 in Science & Mathematics Engineering

11 answers

Actually it is already being researched. Try searching for quantum computing or spintronics
To put it simply - quantum computing aims to use quantum states (which can go up to 32) of a sub-atomic particle.
This answers your question about using a higher base - higher the number of quantum states that you can "read" or "set" higher is the base of your "quantum computer".
I hope that was clear.

2007-06-14 07:14:18 · answer #1 · answered by Hawk 2 · 0 0

Anything that can be done in a higher base can be done in binary

Binary is very easy electrically

But I must admit that I've often wondered about a base 3 computer for use in neural simulation where you could have an excitation response (+1), an inhibitory response (-1) and an indifferent response (0). These, of course, can be - and are - handled numerically as positive and negative weights in neural nets, but it just seems as though a 3-state computer would have an edge.


And, hey, don't knock analog computing. Analog computers, though lacking in precision, produce their solutions at nearly the speed of light and are extremely good at problems involving differential equations, something that digital computers require a lot of time to solve.

At Penn State we had the best of both worlds with a hybrid computer, an analog computer that was connected to a digital computer. The digital computer could adjust the input settings for the analog, read its outputs, make logic decisions, produce human friendly outputs, record results, readjust the analog inputs and go for another run. It was an extremely powerful device. Don't know if it's still there or not.

2007-06-14 08:01:37 · answer #2 · answered by dogsafire 7 · 1 0

There is, it is called "analog" computing, and it has been around long before digital.

Analog systems are susceptible to "noise" and other stray signals (If the computer gets a pulse of, say, 3.4487 volts, was that supposed to be a 3 volt pulse, or a 4 volt pulse?). All electrical signals vary somewhat; none stay constant.

With digital, either a signal is there (binary 1) or it isn't there(binary 0). So a 3.475 volt pulse is just as much a binary 1, as a 4.0203 volt pulse is (they are both interpreted by the computer as a binary 1).

2007-06-14 07:15:48 · answer #3 · answered by Randy G 7 · 0 0

1) Signal or no signal, high or low voltage, on or off, heads or tails... binary. How do create a transistor that is on, off, or something else determined?

2) Error correction. Error deetction in a binary stream is trivial. Error correction given not too high a proability of error is also easy and efficient.

3) The limit of your proposal is an infinite number of channels... analog. Analog is sloppy to implement and imprecise in results.

4) If you want more binary bandwidth, multiplex.

5) Claude Sahnnon showed that the information capacity of a channel is proportional to its frequency. Analog or digital, that is all you can get. Everything else is error correction - and that means binary.

6) You can FedEx 100 Blu-Ray DvDs overnight. Hardly larger than a lunchbox. It's only sneaker-net, but it did deliver theoretical (200 GB)(100) = 20 Terabytes of information. How long would that take to transmit over Cable?

2007-06-14 07:20:40 · answer #4 · answered by Uncle Al 5 · 0 0

Yes, it has been done. There have been computers built in base 10, primarily to produce exact mathematics for business purposes and avoid the errors that can crop up because of the differences between base 2 and base 10 especially in division because repeating fractions 0.123123... in decimal may be exact in binary.
But it is felt that this is achieved by throwing away about third of the capabilty of the machine (the numbers 10-15 in hex A-F)
And while multilevel communications methods have been used, exactly how are you going to store green31 red12 and when you want to add the numbers, how will you do that? If your answer is to convert to binary or decimal, there goes all your speed.

2007-06-14 07:17:58 · answer #5 · answered by Mike1942f 7 · 0 0

Digital (binary) technology enjoys noise immunity. Because there are only two legal states, on and off, we can design a nice safety band between them and still operate reliably at a very low voltage.

In order to detect and discriminate a continuum of signal levels, the pass band for each discrete level must shrink, and that makes the whole system that much more susceptible to noise.

Even more severely, solid state electronics is very sensitive to thermal variations. In digital mode (completely OFF or saturated ON), the components do not perform differently with temperature variation. In linear mode, which is what you need to detect values between OFF and ON, solid state performance varies a whole lot with temperature variations.

To compare solid state to an car, OFF is like the the car is parked with the engine off and ON is like the car is moving at full speed. It is easy to tell the difference from a block away. If you want to tell the difference between the car moving 50MPH and 60MPH you need more sophisticated measurements. Even worse, there will be times the car was supposed to be moving at 60MPH but it's not going that fast because of engine problems or external factors. The data becomes much less reliable.

2007-06-14 09:44:39 · answer #6 · answered by semdot 4 · 0 0

Actually one of the first telegraph systems (English) used an analog voltage. A knob with letters at one end sent a voltage to a meter calibrated in letters at the other end. It was not successful. There is also a story that in the early days of digital computers the Russians built a three-state machine because their first experiments with core memories produced memories that had three states instead of two. So, they just built the whole machine that way.

2007-06-14 18:29:16 · answer #7 · answered by ZORCH 6 · 0 0

That's a good question.
As to the fiber optic question, they can send multiply signals down a fiber optic line that "piggy back" on one another like you say. This is called multiplexing. It still has to be broken down to base 2 binary at the processor though.

2007-06-14 07:19:51 · answer #8 · answered by Anonymous · 0 0

They both sound like great computers so if you want to help your aunt out, go ahead and switch. I love my laptop but I would rather a desktop so that I can have more memory and all that good stuff.

2016-05-20 03:01:54 · answer #9 · answered by leslee 3 · 0 0

It exists, but so much binary programming exists that it is the standard.

2007-06-14 07:11:17 · answer #10 · answered by Anonymous · 0 0

fedest.com, questions and answers