Future of computers

[dj4uk]

Where does everyone see it going?

I've heard about optocomputers, quantum computers, biocomputers - what do you think will take over from the silicon chip when it reaches its limits in approx. 2020?

DJ

[Demosthenes]

One thing that I was told about, but was told would never make it far, was chemical computing. They created a chemical in some University in California (cannot remember if it was Stanford or someone else) that was like a big circle. This circle chemical travelled along other chemicals which formed a line. When electricity or some other stimulus was applied, the circle chemical would go to one end (labeled zero) when the stimulus was removed, or perhaps a different stimulus applied it would go to the other end (labeled one). This gives you the binary sequence needed. It is good in theory, but hard to implement on large scale.

Another idea I liked was using DNA. This theory would never make it either, at least in the speakers view. But, if DNA has two strands, this would allow for powerful theoretical multiprocessing. The example they used was from Gataca when they take a sample of DNA and it comes up with an identity in a millisecond. What was proposed was a bunch of nano machines in bunches of threes. Each bunch of three would have a different combination of Adenine, Thymine, Guanine, and Cytocine. Since there are 4*4*4 = 4^3 = 64 arrangements of the nucleotides allowing repetitions, you would need 64 different types of nano-machine bunches. Then, you would take a cell, throw in a load of these 64 different nano-machines, and then they would align themselves faster than any human could. When an electric signal was sent through them a unique answer would be the result. It would not be in a millisecond, but it was said that current computers could do this in less than 5 seconds, which is amazing considering the length of even very small DNA strands.

Third: using "programmed" cells. Since nature has evolved over billions of years cells that grow up to do specific tasks, humans plan to this in a century. The military is very interested in this as you will see later. Basically, you could program a bunch of cells to create a table. Then you would need a mound of sugar and some water for food, a little time, and the cells would produce a table. It sounds like a dream, but it probably could be done if we knew how to program the cells. The military could obviously take these cells, drop them in the middle of nowhere with some sugar and water and in a few months, they could have a whole town, airstrip, or just a tank.

Lastly, nanotubes. Also, good in theory but extremely hard to manipulate. With the advent of C60 or whatever, nanotubes have been an interesting development. Being so small, I do not think we will be able to accurately manipulate them in the near future, let along create a real circuit. But, computing on the nano-level is full of infinite possibilities.

I think someone will develop a better fabrication/manufacturing process for silicon before any of this happens. Also, for specific hardware devices, the speaker really pushed the importance of FPGA (Field Programmable Gate Arrays) which could are chips that can be programmed after fabrication. Since they are specific chips (like cell phone chips) they can execute their commands hundreds of times faster than other chips.

Respectfully,

Demosthenes

[reboot]

The quantum computer theory is already in practice. We get 4 voltages (accepted standards) with variations (RAM for example at 3.3v) from a power supply. They are +12v -12v +5v and -5v. Applying the positive versus negative voltage in silicon, we then get 0 (off) 1 (on +5), 2(on -5), 3(on +12), and 4(on -12). In practice this is impossible, and voltages are lowered to acceptable limits, getting something like this: 0 (off), 1 (on +5v), 2 (on +3v), and 3 (on +~1.5v or +~7.5v). Getting the chip to recognize a voltage difference as a state (not just on/off) is the hard part.

[archie]

Don't know ... 20 years is a long time considering how quickly things change when it comes to computing ... the developments that will occur in the next 20yrs will surpass the developments that occured in the last 20 [on a logarithmic if you know what I mean] ... and a lot as happened since '81 ... even in just the last 5 yrs.
A breakthrough will occur when a way of eliminating heat, or at the very least reducing it, is found.

[audiyoda]

Originally posted by reboot The quantum computer theory is already in practice. We get 4 voltages (accepted standards) with variations (RAM for example at 3.3v) from a power supply. They are +12v -12v +5v and -5v. Applying the positive versus negative voltage in silicon, we then get 0 (off) 1 (on +5), 2(on -5), 3(on +12), and 4(on -12). In practice this is impossible, and voltages are lowered to acceptable limits, getting something like this: 0 (off), 1 (on +5v), 2 (on +3v), and 3 (on +~1.5v or +~7.5v). Getting the chip to recognize a voltage difference as a state (not just on/off) is the hard part.

Thanx 'boot....now I've got one hell of a headache.

-Craig

[Charliey]

Theres is actually another type of computer theyre developing, molecular computers.

They use very small arangements of atoms to make transistors, eg they are only 3 or 4 atoms long, and are basically the same as current technology, except much much small, and there fore require less power and make less heat.

Now the problems with it is they still havent figured out how to make all the different kinds of logic gates and such with molecules.

Its a bit different from quantum computers because that works on the basis of controling the spin and angle of different particles.

[Demosthenes]

Molecules would be next to impossible to arrange. They are having enough problems with nanotubes. I saw pictures of their only successful attempt to make a gate and it was sad when compared to what they can do with photolithography. So, we will require much more work before we can get down to nano-technology, let alone molecules. I do not see why they have this trouble, though. IBM spelled their company name in Xenon atoms a while ago, and technology has certainly progressed since then.

As for quantum computers, my question is: how would you program them?

Respectfully,

Demosthenes

[Frey Grimrod]

Ahh this is all so interesting there already working twards all of these and more its gonna be fun to see wich way prevails. In the end I think it wont necessarily be the best technology but rather the best marketed look at the automobile thats how the combustion engine took over and won our hearts good marketing to the mass's by Henry Ford. I personally feel direct biological electronic implant techknowledgy is the most interesting to me. Granted not new form of computing but rather a new way of interacting with the machine. Some are starting to make good stabs into that field aswell but this will mostlikely never become mainstream or at least not for a few generations after it has been proved safe easy and effective so that religeous moral issues etc. can be worked out. wow look at me off topic as usual :-P neat to see what everyone else is thinking.

[Demosthenes]

If by "biological electronic implant technology" you mean integrated chips in people's brains and such, that would sound scary to me. Do not get me wrong, I probably do not fear it for the reasons you are thinking. That concept, written in so many sci-fi novels (This Alien Shore by C.S. Friedman probably the best modern example) is quite fascinating, and I probably would want to be the first in line for such an operation, even if it were experimental. But, as illustrated by Friedman, when technology gets widely adopted, it is widely abused. The thought of someone hacking my computer is disturbing enough, but I can deal with it. The worst they could do is destroy a piece of software or hardware. When that hardware or software is directly linked to lets say your optic nerves, then I get queezy. Hacking people was not in the brochure. Just think of all the "computer literate" people who get taken by internet scams and amateur hackers. So, while I think it would be the neatest thing since pre-packaged peanut butter and jelly sandwich slices, it would be a dangerous investment. If by "biological electronic implant technology" you do not mean integrated chips in people's brains, and such, then my ramblings mean nothing.

Respectfully,

Demosthenes

[Colonel Sanders]

I voted optic, but I think it will be part optic and part quantum. Twenty yers is a LONG time in the computer world, like we've reached 2GHz in twenty years(I think 186s came out in the 80s). Notice taht back then, computers were slowly getting better in increments of rougly 30MHz, now they are improving by 100MHz. Eventually GHz will look like MHz, that will be computing power!

So, optic is fast(speed of light, cant beat that), combined with the technology of quantum developed in 2010, you have one awesome PC.

Logan