Computing with flowers
Can you do computing with plants?
Back in the 18th century, the Swedish botanist Carolinus Linneaus created the Horologium Florae, or "sundial of plants." It's described on this web site:
It consisted of flowers that opened or closed at specific times every day. For example, morning glory is appropriately named for its tendency to open in the very early morning. The plants were arranged by the hours that their flowers opened or closed and was laid out like a clock.
Linnaeus studied the opening and closing times to design his "sundial of plants." The daylily closes between 7 p.m. and 8 p.m. Cat's ear opens at 6 a.m. and closes between 4 and 5 p.m. So you see, if you study the habits of enough plants, most hours can be accounted for by such a natural timepiece.
Reading this got me to thinking: Wouldn't it be cool to create a computational logic gate out of flowers? You could rig several different flowers with a set of lights, and set of sensors to determine whether the flowers were open or shut. Under the proper lighting a flower would open (or close, if it were a night flower). And when the flower opened (or closed) it would trigger a sensor that would, in turn, control the lights on a different set of flowers. You could thus set up sets of flowers that would open or shut other sets of flowers, and vice versa.
Seems to me that you could thus pretty easily create the basic logic switches that drive computer chips -- like AND, OR, NOR, or XOR switches. Of course, given how slowly flowers open and close, they'd be the most glacial computer processors on the planet. You could set up a circuit to, say, add two binary numbers -- and then sit back and watch as it takes, like, three hours for the flowers to open and close enough times to do the calculation.
But that'd be the cool part about it! As my friend Greg said when I told him this idea, "it's like a crazy version of the Clock of the Long Now" -- a clock that a bunch of geeks are building that will tick once a year, to remind us humans of how long time is, and how brief our lives actually are. The planet "thinks" awfully slowly, which is precisely what a circuit made of flowers would also illustrate.
It'd also potentially generate some weirdly odd results. After all, computer circuits are designed to be precise and regular. Live things like flowers aren't -- they might open or close unpredictably. A flower computer would thus occasionally produce some wonderfully cock-eyed results. Even better, imagine what would happen when the flowers began to pollinate and grow and spread -- "growing" new switches in the circuit and producing new logic that the planter/builder didn't intend.
Damn, now I wish I had a back yard.
(Thanks to Greg for pointing this one out!)
Posted by Clive Thompson at August 05, 2003 12:20 AM
| TrackBack
I know, the Horologium Florae rocks the house!
Good point about the relatively openness -- or closedness -- of flowers. I guess you'd have to treat each flower like a fuzzy-logic switch, determining its state as "mostly" open or "mostly" closed. But obviously that, too, would create extra weirdness and errors in the system. In a way, it would produce even cooler results, because if you set up an 8-bit binary adder and told it to add 134 and 289, you wouldn't necessary get the mathematical total -- you'd get what the flowers think the answer is, heh.
Ooooh man! What a superb idea! As for determining whether a flower is 'open' or 'closed' I think the sensor array best suited would be optical. Have a digital image taken of the bed as a whole, then interpreted by some software running on an actual computer. Based on the analysis, the computer could then control the associated lights.
It'd be like having the hyperactive child flitting around the ponderous master, waiting to do its bidding. :)
That would make one cool photo-essay and time-lapse video tho'. Could call it "The Nature of Computing" ;)
K
