H2O/IP

AltImage writes "This interesting project uses water as an organic network between two computers. It analyzes the color of each pixel and 'prints' out pulses to the electronically controlled water valve - a different pulse pattern depending on the color of the pixel on screen."

Comments

Water isn't organic..

[Anonymous Coward]

damnit.

Re:Water isn't organic..

[Anonymous Coward]

and it is ip/h2o, not h2o/ip...

How refreshing.

[jericho4.0]

This is a cool project. It reminded me of this project (from this /. posting. A guy building logic gates with water flow.

There was another link I can't find anymore to a lab moving microscopic drops of water around on a sillicon substrate really fast. The target apps are in biochemistry, but iirc the design used the liquid to do some logic, also.

Re:Time

[panurge]

No, because droplets can follow one another with more than one in the channel at the time. They only need to be far enough apart to be sure they are distinguishable.

If we assume that drops need to start off at least 10mm apart, using the high school classical equation t = sqrt(2s/g), we get about .045s or around 20 drops per second. (They will be physically much further apart at the bottom, of course, but the same distance apart in time.)

Now assume we use classical serial communication, 1 start bit, one stop bit, one parity bit. That's 11 bits to a byte, or very roughly 2 bytes per second. The problem is mainly one of error correction. There is no back channel, so any errors cannot be corrected as there is no retransmit request. This is definitely not related to TCP/IP which was intended to be a robust protocol. It's just the equivalent of Morse code. Even so, at about 100 bytes per minute, and with the opportunity for compression, the transmission rate is about as fast as an ordinary Morse operator.