why binary?

[Dr_Doggy]

as we all know computers use digital, 0-5v, 0 being a 0 and 5v being a 1,

so why don't we use logic where on ONE BIT you can have 1-4, ie, 0v=00 1v=01 2v=10 3V=11

or some other similar method, it would compress data quite a bit and free up wires?

 

[ARandomOWl]

That device is known as an Digital-Analogue converter or DAC.

 

[JimB]

as we all know computers use digital, 0-5v, 0 being a 0 and 5v being a 1,

so why don't we use logic where on ONE BIT you can have 1-4, ie, 0v=00 1v=01 2v=10 3V=11

or some other similar method, it would compress data quite a bit and free up wires?

You are completely screwed up in your head.

Bit = Binary Digit
A bit can be 0 or it can be 1, that is the whole point of DIGTAL devices, things are either switched on or off.

So these ANALOGUE voltage levels which you are proposing, how would you represent them in a DIGITAL computer?

0v could be 00
1v could be 01
2v could be 10
3v could be 11

er....

there is a bit of a circular argument here, like a cat chasing its tail.

JimB

 

[Dr_Doggy]

that's what im saying, why not have analog computers ?

 

[3v0]

You post should have ask why digital ?

There are analog computers. We had one at the U I attended. It was series of blocks you connected with patch cords.

There may be some modern ones that allow you to program the connections digitally.

 

[crutschow]

Multiple state digital logic is not silly, and has actually been proposed as a way to minimize connections (see this for example). The reason it is not normally used is that it significantly complicates the circuits, since they are now required to detect more than two levels (high or low). This added complexity generally means it's easier to add more lines to carry the signal than add more complex circuitry.

 

[misterT]

as we all know computers use digital, 0-5v, 0 being a 0 and 5v being a 1,

so why don't we use logic where on ONE BIT you can have 1-4, ie, 0v=00 1v=01 2v=10 3V=11

or some other similar method, it would compress data quite a bit and free up wires?

How would you make a transition from 00 to 10 without going through state 01?

In fiber-optics it is possible to send data in different wavelengths simultaneously. The same can be done in an electrical wire.. and is definitely done in radio communications.

Wavelength-division multiplexing - Wikipedia, the free encyclopedia

 

[crutschow]

There are analog computers. We had one at the U I attended. It was series of blocks you connected with patch cords.

There may be some modern ones that allow you to program the connections digitally.

Way back when they did build hybrid analog-digital computers for control loop simulation, since the digital computers weren't powerful enough at the time to do complex feedback control loop calculations in any reasonable time frame. The digital computer controlled the connections, the state of the computer, and processed the data output, while the analog portion performed the control loop simulation.

But I don't think they have those anymore.

 

[crutschow]

How would you make a transition from 00 to 10 without going through state 01?

You go through the state but you don't detect the final state until the transistion has settled. It's like clocked logic.

 

[Sceadwian]

Ternary logic can be implemented using tri-state I/O lines that have pullups available.

Doggy there are analog computers, there have been since long before digital computer existed, and they have all the problems of an analog system. Since a single bit can have four states there are for more chances for that state to be misread as an adjacent one, where as binary logic has only two possible states EVER. You could read a binary dataset and write it an infinite number of times and have no data corruption, where as with additional states the data can be in with analog computer the higher the chance of a read error. That's not even the core of the reason though. Transistors can be VERY easily made to attain a binary state where making them achieve solid reliable, repeatable, and enduring additional states the reliability drops very quickly and the devices become more prone to both noise and thermal effects.

 

[crutschow]

As Sceadwian noted, the main reason to use two-state binary circuits is that it's very easy to build circuits where the transistors are either on or off. The tolerances required to build such switches are wide, making it relatively easy to build very large, reliable digital circuits. All you need to do is make sure that the transistor is either on or off, not what fraction is it on or off.

 

[CafeLogic]

People are stupid. I, for one, can't remember where I put my car keys half the time. You think I would be able to keep track of enough things to make any kind of decent program in a multistate computer? hell no! All of this magnificent progress you see is because some genius decided to break everything down to 2 values, so that even simple minded folk like myself could contribute. Oh, and there is also noise immunity, that's kind of important. It's the reason that, much to the dismay of Sony Music, even after that MP3 gets copied 800 million times, it still sounds exactly the same. Try that with your cassette tape.

But just so you know you weren't the first person to think of that. This is how gigabit ethernet gets all that information through those little wires. Each voltage level is one of 5 values. 10GBASE-T uses 16 values. If that wasn't enough, remember back in 1992, when you had to wait forever for your dirty pics to download at 2400 bps? Well, if it wasn't for the magic of Pulse-amplitude modulation, that transmission speed would be limited to 300bps.

 

[dknguyen]

For the complexity reasons as previously stated.

It tends to be used only where availab ebandwidth is very limited or can't be easily increased by just adding more wires like RF or communications as someone mentioned earlier. In RF they can have multi-level signalling to make better use of the very limited radio spectrum to increase data rates. Transmitting and receiving circuits convert it back to/from binary though because its easier to work with.

 

[sangeo949]

hai, the actual reason behind this is, each and every electronic devices have a linear switching chara: ie, no device is found to switch in a step chara: because of this the component does not switch exactly on a particular voltage so we need a range of voltage to specify a value (usually 0-2.4 as zero and 3above as 1)