Better programing technicques...
well there's a thought! :-D
Boffins at the Incheon National University have made what they claim to be a breakthrough in computing: a programmable processor which uses DNA, rather than electronics, to perform its computation. The Microfluidic Processing Unit (MPU) developed by the team is, it claimed, a step forward in simplifying DNA computing – a …
The prototype MPU is, admittedly, basic, offering only AND, OR, XOR, and NOT operations
You are aware that every digital computing device is implemented with the AND, OR and NOT functions, I assume? Indeed these are the only functions required for digital computing.
With sufficient gates offering those functions you can get arbitrary complexity.
XOR is not a base function in Boolean algebra -> Y = (A AND NOT B) OR (B AND NOT A) where Y is the output, A and B are the inputs.
I think the point is that so far all they can create is a few logic gates. This research demonstrates that cascading calculations through multiple gates is possible, but it looks as though there would be serious constraints on how far that could be taken. Some of those may be just engineering problems - it's slow, results can be either read or passed to another gate but not both, and various other things. But the fundamental nature of the reactions means that the DNS strands being used change in each step. For the AND gate, you start off with 20 base pairs in each input, and get 40 out at the end. After passing through a second gate, you now have 80 base pairs in the output (figure 8 in the paper). This would seem to be a pretty big problem, literally, if you want to have billions of gates operating together.
So it's not so much that it's only a basic computer because it only offers a few logic gates, but because it only offers those gates with no clear path for how you could build a functional computer out of them. When the transistor was first invented, it was pretty obvious that you could build huge networks of them, even without knowing exactly which direction the technology to do so would end up going. In this case, it's not at all clear that it will be possible to build such networks at all. The hope is there, but it's not just a case of figuring out how to do it, but figuring out if it's possible at all.
Biting the hand that feeds IT © 1998–2021