Computer Scientists Build Computer Using Swarms of Crabs - Technology Review(technologyreview.com)
technologyreview.com
Computer Scientists Build Computer Using Swarms of Crabs - Technology Review
http://www.technologyreview.com/blog/arxiv/27730/?ref=rss
6 comments
It takes the author to the end of the piece to admit that no, in fact, they did not build a computer using swarms of crabs. Right after he admits it's false he, he asserts it is true again...really...link baity bad writing for an MIT blog.
"To be fair, the results were mixed. While Gunji and co found they could build a decent OR gate using soldier crabs, their AND-gate was much less reliable.
However, it's early days and they say it may be possible to produce better results by making conditions inside the computer more crab-friendly. (No crabs were harmed in the making of their computer, say Gunji and co.)
So there you have it--a computer in which the information carriers are swarming balls of soldier crabs."
"To be fair, the results were mixed. While Gunji and co found they could build a decent OR gate using soldier crabs, their AND-gate was much less reliable.
However, it's early days and they say it may be possible to produce better results by making conditions inside the computer more crab-friendly. (No crabs were harmed in the making of their computer, say Gunji and co.)
So there you have it--a computer in which the information carriers are swarming balls of soldier crabs."
I know what I'll be building next time I visit the sea side. No more sand castles...
Somewhere Charlie Stross is smiling widely.
"The first is theoretical--ordinary computers are hugely energy inefficient--some eight orders of magnitude worse than is theoretically possible. "
Eh? Is this saying that ordinary computers are eight orders of magnitude more energy-inefficient than the maximum possble inefficiency?
Oh, wait, they mean the other possibility.
Also, as others mention, they only build two gates, and one of them doesn't work very well.
Eh? Is this saying that ordinary computers are eight orders of magnitude more energy-inefficient than the maximum possble inefficiency?
Oh, wait, they mean the other possibility.
Also, as others mention, they only build two gates, and one of them doesn't work very well.
I'm waiting for the NAND or NOR gate...
The NOT gate might be difficult to build ;)
They claim to have built it already, and given that the AND gate is actually a tripartite output of a∧¬b, a∧b, ¬a∧b the NOT seems to be trivial.
Note: this is not level based logic, but billiard ball logic. If I understood that better I'd write a concise, illuminating explanation of why ¬a∧¬b is missing in the AND gate's outputs.
http://en.wikipedia.org/wiki/Billiard_ball_computer
Note: this is not level based logic, but billiard ball logic. If I understood that better I'd write a concise, illuminating explanation of why ¬a∧¬b is missing in the AND gate's outputs.
http://en.wikipedia.org/wiki/Billiard_ball_computer
Uh, I think you're overthinking this. What is the output of a NOT crab-gate when there are no crabs in the input?
In a billiard-ball model one would construct a NOT/inverter with a crab source that normally flows straight to the output but can be misdirected into a drain by crabs coming in from the input. Of course you need a crab source for this, but this is little different from electronic transistor-based computers, where the inverse of Vss is Vdd and the Vdd comes from the power supply.
Right, yes, but the (semi-serious) point is that we don't have a crab grid in the same way we have a power grid, so actually implementing this (which is what the article is about, rather than just theory) might be hard. For the crab computer to "run" arbitrary computations it would need an arbitrarily large source of crabs.
A functional crab computer would probably have to be built from delay insensitive logic[1]. The outputs of the gates would be of the form X and !X, where the signal (crab cluster) would appear at the appropriate output.
With regards to crab supply, I think the sufficient study of reversible computing, and in particular their charge conservation properties, would address that.
You'll still have to have a mechanism to spray bits of dead fish around to feed the thing.
--
[1] http://en.wikipedia.org/wiki/Delay_insensitive_circuit
[2] http://en.wikipedia.org/wiki/Reversible_computing
With regards to crab supply, I think the sufficient study of reversible computing, and in particular their charge conservation properties, would address that.
You'll still have to have a mechanism to spray bits of dead fish around to feed the thing.
--
[1] http://en.wikipedia.org/wiki/Delay_insensitive_circuit
[2] http://en.wikipedia.org/wiki/Reversible_computing
Still, building logic gates using crabs, very good. :)