Anything's possible. It's just gonna be really fucking hard to pull off. The reason we're using binary is because it's so damn simple. On/off. That's why we'll never understand women.
Name:
Anonymous2006-06-01 19:26
Theoretically, if we ever start using Light instead of Electricity, we could easily do > binary. 5nary could be Black (0), Red (1), Green (2), Blue (3), White (4).
But practically, i have no idea.
Name:
Anonymous2006-06-01 20:29
>>5
i know someone who tried to build some kind of light-cpu (or at least he experimented with it), but he said it wouldn't work because it needs something that changes the wavelenght or something instantly, when you send the light through it (like some lens or shit). makes sense, because if you don't have such a thing you would have to generate new light everywhere, which would slow it down.
Name:
Anonymous2006-06-01 21:41
hmm there are more than two quantum states, correct?
Name:
Anonymous2006-06-01 23:16
binary is "yes/no"
does that make Trinary "yes/no/maybe"?
does that mean a bit will now be called a tit?
:D
set back gender equality in computing another 10 years!
Name:
Anonymous2006-06-02 0:37
Fuck trinary computing. Why not just have straight up analog computing whereby each "bit" is a measure of energy that can be any one of an infinite number of states. This means such analog computers can have notions of infinity, and inherent fuzzy logic, and a whole bunch of other interesting stuff.
Oh wait. They already got that. It's called the human brain.
Name:
Anonymous2006-06-02 0:38
Furthermore, some of the data storage mechanisms of EAROMs in old arcade boards have three states, a on state, an off state, and an erased or unwritten state.
Name:
Anonymous2006-06-02 4:08
There's a dialect of INTERCAL that uses trinary.
Name:
Anonymous2006-06-02 4:27
Ternary is the most efficient way to represent data (because it's the closest integer to e). It just happens that binary digits are the simplest to handle algrebraically, even despite stupidity like one's and two's complement.
And ternary computing is very possible; a number of ternary computers have existed.
By the way, to the guy who said we should go with analog computing, I regret to say that analog computers are rather useless. They're only useful in a few niche domains.
The human brain isn't exactly a true analog machine either. Either an axon fires, or it doesn't. That's not to say it's digital either, far from it (different activation states, neurohormones, sensitivity at a particular synapse, etc). It's pretty weird in there.
Name:
Anonymous2006-06-02 5:14
The martians in "A Stranger in a Strange Land" counted in trinary. Powers of 3 and all that shit.
Name:
Anonymous2006-06-02 9:13
0, 1, 2, 10, 11, 12, 20, 21, 22, 100 READY OR NOT HERE I COME
Name:
Anonymous2006-06-02 9:30
>>12 Ternary is the most efficient way to represent data (because it's the closest integer to e)
Could you please explain this?
Name:
Anonymous2006-06-02 10:45
yes it's because the state (e^ik)|phi> is the same as |phi> and 3 is closest to e, so yuo get much less errors
Name:
Anonymous2006-06-02 11:54
>>16
The link I provided does a better job than I could hope to do.
I personally think their metric is rather arbitrary, but I'm not a mathematician.
Since all data transmission is electrical impulses, (assuming not net traffic over FO) it is easiest use only two states, easily handled by logic gates and transistors. IT can simply be +5V/0V or +5V/-5V. Quantum computing would allow for trinary computing because you could have "on", "off" and a quantum super-position of the two, both on and off, to make three states. You could try telling the machine to look for more voltage levels as states, but since even the fastest changes in voltage are not QUITE instantaneous, a change in value from lowest to highest voltage will generate a discrete value from another state that it briefly hit along the way.
>>12,18
That was a great read, thanks for the link. Indeed base 3 is more efficient than base 2 (if the hardware were equally as simple), although not being able to easily divide by powers of 2 is a practical drawback (dividing by powers of 2 is more useful than powers of 3).
BTW, two's complement is not stupid!
Name:
Anonymous2006-06-02 20:47 (sage)
I personally think their metric is rather arbitrary, but I'm not a mathematician.
Uou're completely right though. There isn't one optimal base for all (or even most) applications.