Trinary Code

[MyndFyre]

EDIT: I will try to explain this one last time. It won't have any rule for it. The MAJORITY of prime numbers will end with a 5 or a 7. I never said anything about all of them. If this post is too bland for you, and my theory is too, just simply disreguard it.

Um...if a number ends in 5, then it's divisible by 5 and therefore not prime.

In a number system with 12 specific digits (let's say they go up to B), he's saying that most prime numbers will end in 5 or 7.  Like so:
(base)
10    12
5      5
7      7
11    B
13    11
17    15
19    17
23    1B
29    25
31    27
37    31
41    35
43    37

Not that that's significant in any way...   by adding 2 to the base, you're eliminating only one (for that SMALL subset of primes) and you're gaining one ending.  OOOOOHHHHH.

[dxoigmn]

In a number system with 12 specific digits (let's say they go up to B), he's saying that most prime numbers will end in 5 or 7.

Bah!  That's what I get for not reading the whole thread

[idoL]

How do you nay-sayers know this wouldn't work? Can you show me your research for proof and provide back up claim from other sources? If not then stop saying it's stupid/waste of time.

[dxoigmn]

How do you nay-sayers know this wouldn't work? Can you show me your research for proof and provide back up claim from other sources? If not then stop saying it's stupid/waste of time.

I don't think the burden of proof is of the nay-sayers.  Rather, it is the burden of whomever came up with the ideas to convince everyone else.

[MetaL MilitiA]

Not at all, binary is soo damn repetitive, trinary would have a lot more combinations, meaning less code for more stuff. Please back yourself up next time.

OK.

A binary logic gate can be built with two transistors, and the transistors can be made very inexact and small because they only have to be "on" or "off". Trinary logic would need to use more states than just on or off. Such a transistor, calibrated for analog use, is very much larger. Hence the waste of space.

Furthermore, binary logic can be made faster because the target output level is at the limit of the range - maximum or minimum. You do not have to worry about an overshoot, and so there's much less of a stabilization time for the output. Think bang-bang regulators - why are they so fast? With trinary logic, you'd have a target output level that you'd want to reach, there'd be overshoot, and all in all the input to output delay of any gate would be larger than for binary logic. Hence way slower.

This does not prove that trinary code would be slower, all it proves is my idea for the representation of the data is poor and would be slower. It would not effect the way the math with trinary code would work. I guess trinary code would not be possible with the current technology we have, but that does NOT mean that it would be slower than trinary in any possible way. THat you for saving me an argument MyndFyre. dxoigmn, I wasn't trying to convince other people, as there is no point in trying to convince somebody 1+1=2. Please, if you discuss this further, try to actually prove the idea of trinary code wrong. Or, you can even add to the idea, or just even ask questions.

[MyndFyre]

This does not prove that trinary code would be slower, all it proves is my idea for the representation of the data is poor and would be slower. It would not effect the way the math with trinary code would work. I guess trinary code would not be possible with the current technology we have, but that does NOT mean that it would be slower than trinary in any possible way. THat you for saving me an argument MyndFyre. dxoigmn, I wasn't trying to convince other people, as there is no point in trying to convince somebody 1+1=2. Please, if you discuss this further, try to actually prove the idea of trinary code wrong. Or, you can even add to the idea, or just even ask questions.

You are correct; however, the medium by which the data would be transfered is inextricably tied to the system itself.

As Adron and I pointed out, the reason that binary is so efficient with our current technology is that there are two states: on or off.  Two states doesn't inherently make it efficient, but the fact is we do not need to be remarkably precise with voltages.  If we were going to use electricity, we'd need to increase the range of acceptable voltages or decrease the acceptable "miss".  The former will increase heat and degredation in the components; the latter, more misses.

[R.a.B.B.i.T]

There is, sort of, already a kind of FORM of trinary, which it actually isn't.

I forget what it's used in, I'll have to ask my teacher again when I go back to school, but I do remember the states:
0 - low
1 - high
x - high resistance, so 0 or 1 doesn't matter

[Adron]

This does not prove that trinary code would be slower, all it proves is my idea for the representation of the data is poor and would be slower. It would not effect the way the math with trinary code would work. I guess trinary code would not be possible with the current technology we have, but that does NOT mean that it would be slower than trinary in any possible way. THat you for saving me an argument MyndFyre. dxoigmn, I wasn't trying to convince other people, as there is no point in trying to convince somebody 1+1=2. Please, if you discuss this further, try to actually prove the idea of trinary code wrong. Or, you can even add to the idea, or just even ask questions.

With your logic, infini code would be the best one. In this code, we don't deal with multiple digits. Every number could be represented with a single symbol. This code would be extremely fast of course. Now I have added to the idea, brought it to the far end of additions, from which there can be no improvements. The part of discussion that doesn't cover implementation is completely exhausted.

However, in reality, implementation is what matters. You cannot measure speed of trinary vs binary vs infinary without considering implementations. Implementation is all that matters.

[Adron]

There is, sort of, already a kind of FORM of trinary, which it actually isn't.

I forget what it's used in, I'll have to ask my teacher again when I go back to school, but I do remember the states:
0 - low
1 - high
x - high resistance, so 0 or 1 doesn't matter

That's called tristate. It's used when you hook outputs up to a bus. All inactive devices tristate their outputs.

[MetaL MilitiA]

Then lets just discuss ways we could make it work now. We need to think of something else other than an electric current (Notice I didn't use the inaccurate word of electricity, when eletricity actually moves slower than maple syrup.). Maybe Fiber Optic, since there are different frequencies of lights too. I'm not to familular with that though.

EDIT: Now that I think of it, this could work in a great way to, as I'm sure there is a way to make a certain frequency of light cut off into a different channel using a certain type of prism, so it would be as simple as possible still.

Diagram:

Code Select Expand


                 /
------------(---
                 \
(=prism used to seperate the different frequencies.
/=one frequency would use this path
\=another frequency would use this path
--- after the (=a third frequencie would use this path.

In theory, this would be the fastest way to do it, as long as when the light bends it does not delay the information.

[dxoigmn]

Then lets just discuss ways we could make it work now.

Quatum computing?

[MetaL MilitiA]

I havn't really looked too much into Quantum Physics, could you give a breif summery on how it works?

[MetaL MilitiA]

To answer your last question there since I'm too lazy to sign up to the vL forums.
Quantum Physics is a way trying to describe the universe on a subatomic level. It is far opposite to regular physics since it doesn't have exact answers. The Uncertainity property states that the more your know about something the less you know about something else. Like if you know an atoms position you can only guess at the speed its going and vise-versa. But I think what the kid is talking about by Quantum Computing is building a computer from the atom up, using quanti (small packets of energy) of enery to run it.
BTW I love that topic, really makes you think.

He gave a pretty breif summary, which makes me understand the majority of it. Anybody like to add?

[Adron]

In theory, this would be the fastest way to do it, as long as when the light bends it does not delay the information.

All you have *really* accomplished there is recovering the binary code to do some work. Multiplexing of signals in fibres already exists. You can transmit data on several different wavelengths at once to gain a higher transfer speed. That's a matter of encoding information for transfer though, not for actual computing. And you're still using binary code, just transmitting more than one binary digit at once on separate channels in the same fibre.

[MyndFyre]

In theory, this would be the fastest way to do it, as long as when the light bends it does not delay the information.

All you have *really* accomplished there is recovering the binary code to do some work. Multiplexing of signals in fibres already exists. You can transmit data on several different wavelengths at once to gain a higher transfer speed. That's a matter of encoding information for transfer though, not for actual computing. And you're still using binary code, just transmitting more than one binary digit at once on separate channels in the same fibre.

Vis-a-vis parallel data transfer.

I gave thought to using quantum principles, too, when considering whether or not your trinary system would work.  The best example I could come up with is the "mini-teleporter" that they have not, which they can use to align the charge of an ion a certain way -- specifically, exactly the way that another ion is aligned at an arbitrary distance away.  The practical application of this is not yet known, but they suspect it to be the beginnings of a rudimentary teleportation system, similar to what is seen in Star Trek.

The problem with this is that (currently) you have only a 25% chance of getting it right.  Not really a practical application.  Oh -- and you need a particle accelerator.  Fit THAT into your bedroom.

To also build onto the fiber-channel idea: you COULD calibrate a 3rd light frequency.  For ease of discussion, let's say it's based on intensity, like a monitor pixel (0 is off, 255 is the highest).

With this range, you could say:
0-127 -- considered 0.
128-255 -- considered 1.

If you wanted to get ternary, you'd have to reduce that:
0-85 -- considered 0
86-170 -- considered 1
171-255 -- considered 2

The problem with this is the same problem I described with sounds and electricity -- you need a more fine calibration, and you will end up with more misses on data.  If you're going to go through all that trouble, why not just:

(R) 0-127 -- considered 0
(R) 128-255 -- considered 1
(G) 0-127 -- considered 0
(G) 128-255 -- considered 1

and then combine them for faster, parallel transfer?