Slashdot Mirror
Handshake via the Internet
mattlamb writes "British and American scientists will touch using sensors over the internet. "The implications of the experiment could be vast, said UCL, which describes the event as the world's "first transatlantic handshake over the Internet." " Let the juvenile comedy
commence!
At last, I can be gay over the Internet.
Well, I was, but now I can be more gay. A gay enabler, if you will.
Now I'm the San Francisco treat.
Sampling x number of sensors with y bits of precision 1000 times a second? Um yeah, that's pretty damned high bandwith.
... for just the sensor info of the handshake. If it is more like 1000 per square inch, then we talking even more.
:)
Figure you need a fairly high value of x to determine things like woody / fleshy material, lets say 100 per square inch. (I have no idea if 100 is really reasonable, it may be more like 1000.)
Then you also need some decent resolution of those sensors, say 8 bits worth.
Then we get 100 * 1000 * 8 = 800000 b/s or around 800 kb/s (would that be 781?)
Now, my hand has more than 1 square inch of surface area, so scale appropriatly.
Subtract compression, add TCP overhead, and that's still several megabit at a minimum
And I always thought you only needed 3 TCP packets to make a handshake
there needs to be a (-1 : Predictable) moderation option...
The article glosses over the fact that there are very, very few genuinely practical applications for this because of two insurmountable problems.
For one, our experiments at the time demonstrated that the hand-control idiom suffers from any lag; specifically that delicate manual operations are basically imposible with latency as low as 30ms which rule out things like surgery done this way. Hand dexterity depends on a very large number of relexive immediate movement in response to subtle stimuli like minute vibration of the tool, perceieved resistance, etc.
The second problem is one that operators of such devices very quickly become disoriented, often nauseated, because of the discordance between years of ingrained knowledge of how the world reacts to touch and the lagged/different input such tools provide.
People need to learn /new/ idioms for remote manipulation, not attempt to emulate biological systems. That's the same peoblem AI research has suffered from its inception: the day computers will display intelligence is when researchers start working on computer intelligence instead of trying to simulate human intelligence.
Same goes with tools.
While this might be a geekly thing to do (handshake over the net) and quite a bit neat, it's neither revolutionnary nor interresting in the long run.
-- MG
...Why is it that "scientists have done x over the Internet" is automatically newsworthy? The same demonstration performed over a cable between two adjacent rooms would not have been significantly easier. Stuffing arbitrary data into TCP packets is just not that hard.
Warning: this article may contain humor, sarcasm, parody, and perhaps even irony. Read at your own risk.
How many posts on slashdot is not predictable, esp those opinionated ones.
--- You make things foolproof, and they'll find you a damn fool.
Could this technology be used in some way to create 'feeling enabled' prosthetic limbs? I know there would still be issues with impulse delivery to the brain, etc. Just a thought