So how does one program such a computer?
In order to wrote a DNA program, one would of necessity have to be able to manipulate individual molecules within DNA in a reproducible way, and in huge enough quantities to allow for meaningful computations. And to be useful, the process would have to be rapid.
If we're able to do that, then the medical implications are enormous -- far eclipsing the impact of DNA computing. If we're not able to write programs, then DNA computing is unlikely ever to be more than a stunt.
On a side note, this development adds an interesting twist to the "origin of life" argument.
OK, I know this isn't a scientific article or anything, but has anyone figured out how to write the ROM that will interact with this thing? In other words, what kind of problem did it solve, how was the problem put into the "machine" and how was the result taken from it? If the cellular computer simply responded to a chemical reaction in a known way, I'm not sure of the value.
We are so technically oriented that every little thing sounds like a marvellous breakthrough or something. But there's a huge difference between creating a molecule that acts in a known way and creating a programmable computer. As someone said earlier, the real trick is in generating the DNA in the first place. Given that they can do that, the implications on medicine are far more staggering than the implications on computing.
Shalom.
The automaton's hardware consists of a restriction nuclease and ligase, the software and input are encoded by double-stranded DNA, and programming amounts to choosing appropriate software molecules. Upon mixing solutions containing these components, the automaton processes the input molecule via a cascade of restriction, hybridization and ligation cycles, producing a detectable output molecule that encodes the automaton's final state, and thus the computational result. In our implementation 10^12 automata sharing the same software run independently and in parallel on inputs (which could, in principle, be distinct) in 120 mu l solution at room temperature at a combined rate of 10^9 transitions per second with a transition fidelity greater than 99.8%, consuming less than 10^-10 W.
Not with a 0.2 percent error rate, they won't.