Slashdot Mirror
Future Computers
jethro200 writes "Popsci.com has an interesting story on the up-and-coming silicon replacements, ranging from DNA to a little molecule called thiol to using atoms in a quantum state. Obviously, these are a long way from being your next desktop, but an interesting article nonetheless."
Great minds reading /. who are interested in this article should definitely make their way to Pasadena this summer for the Computing Beyond Silicon Summer School. The dateline for applying has passed, but you can always gatecrash, or monitor the site to read the lecture notes online (they better be available).
This sort of stuff bothers me.
We don't have a 3-5 pound computer sitting in our heads. We have a 3-5 pound brain emulator sitting on our desk.
The point of the computer (originally) was to do complex tasks that took the human brain too much time. It does slave-like replication. It's an emulation of something we can already do, in theory.
Furthermore, the human brain is far from puny. We have 10^15th synapses, which is far more connections than there are genes in our genome, or even stars in the galaxy. 10^15th synapses is an incredibly large number to imagine. A synapse is a neuronal connection. A data transfer point.
I urge the above poster to consider the fact that life and thought have been debated for thousands of years. We *can* be so difficult to figure out.
I thought that this might be interesting for a few of those who don't know much about quantum computing.
The idea is to have a bit that can be a 1 or 0 at the same time. This means that with 50 bits, called qubits, you can represent every number from 1 to 1 trillion, at the same time.
What's really cool, is with this you can use what's called a bogo sort. Imagine a set of cards, that is shuffled. Now to sort them in order, most people would go through 1 by 1 and put some in front and some in back. A bogo sort creates a new universe and then throws the cards into the air. If they land in order, great, else destroy the universe.
All these universes are created at the same time, making it 1 step to sort 52 cards. Like I said, it's interesting.
=================
Unix is very user friendly, it's just picky about who its friends are.
A thiol is not a molecule but a -SH residue so if someone talks of thiols, a whole group of molecules featuring that residue is meant. That's much like alcohols with their -OH residue.
In 1994, Leonard Adleman (the A in RSA) showed that it was possible to solve a particular computational problem using standard molecular biology techniques. His experiment solved an instance of the Directed Hamiltonian Path Problem (also called the Travelling Salesman problem) entirely by manipulating strands of DNA. The problem he solved was only 7 nodes in length, which is easily computable by hand in about 20 minutes, but was a great achievment in molecular computing.
There are two compelling advantages to using molecular biology to solve computational problems. Firstly, DNA has a much greater information density than almost any other media: using DNA it is possible to store data in a trillion times less space than with an electronic computer. Presently, it is possible to contain 10^21 DNA molecules in less than 1 litre of water, (with each molecule encoding potentially 400 bits of information).
Secondly, biological operations performed on DNA are massively parallel. All operations that are executed are performed on each strand of DNA simultaneously.
Adleman's experiment encoded a 7 node Hamiltonian Path Problem. Each node of the graph was encoded as a random 20 base long strand of DNA, and these were randomly annealed into long potential 'paths' through the graph. Paths were selected (and extracted) based on the length of the strand, which nodes were encoded in the strand, and whether the path encoded all nodes. At the end of this selection process, the remaining strand/s should encode the shortest path of the graph.
As can be imagined, this experiment has potentially huge consequences for large computational problems.
The problem remains to make this process reliable.
David Jackson
--
Incorrigible punster - Do not incorrige.