Robo-Arm Signatures Are Legal, Gov't Buys One 154
AndreV writes "It's endlessly comforting to know a recently designed and implemented long-distance robotic signing arm can produce signatures legal in both the US and Canada. The aptly named LongPen replicates the handwriting from a person writing in a remote location — with the unique speed, cadence and pressure of a human pen-stroke. It started as an idea from author Margaret Atwood to help free her from grueling, multi-city, multi-country book tours, but the hard stuff was done by a bunch of Canadian haptic gurus, whose design took into consideration many factors of the human arm and how we write. How it works: from the author-end, data protocols are set up, and the pen pressure is measured on a special tablet. The data streams to the robot, while algorithms smooth out all the missed points. Complex math operations were used to help the mechatronic limb repeat the hand's motions without unnecessary jerking, and programmers had to 'scale time' or 'stretch time' by breaking down the movements, essentially tricking the eyes into thinking the robot is writing fast. It was recently adopted by the Ontario Government to sign official documents. It helps criminals sign books, too."
Re:This is news? It isn't new. (Score:4, Interesting)
Auto signature machines are not the same as long distance signature machines. It's also worth noting that mechanical signature systems are rarely used for sensitive data etc. (they're normally used on cheap merchandise etc. and hand writing experts can tell the difference between the mechanical version and a real signature)
Impact == 0 (Score:4, Interesting)
The value of a signature is its difficulty to replicate. The historical cut off for this has been the talent and prevalence of expert forgers. Having automated forgers is quite irrelevant if they require more investment of time and effort to perform the same replication. (which would clearly be the case for this implementation, at least)
If anything, I would say the problem is that these machines are being underapplied. What they should *really* be used for is to create extremely complicated signatures a human being would not be able to accurately reproduce. Then for the first time in hundreds of years written signatures would become more secure.
(Granted, only until someone develops a machine that can reverse-engineer them, but at that point human-written signatures would have been even less helpful.)
Complicated solutions to easy problems (Score:2, Interesting)
Re:Complicated solutions to easy problems (Score:3, Interesting)
Re:Write once, reproduce more (Score:5, Interesting)
Not only that, how do you know what you're actually signing if you're not there to read it in person?
You don't even need to figure out a way to store and reproduce it. Just through a piece of carbon paper under the document and have a second contract under it, or even just a blank sheet of paper to be filled out later.
Re:The real question (Score:3, Interesting)
The robo-arm doesn't really add a lot or problems to that though (and if someone uses a naive playback attack to forge multiple signatures, the fact that they are too similar should make it easier to successfully deny the signature).
Re:Write once, reproduce more (Score:2, Interesting)
From my experience, when filling forms that use carbon paper, there's no carbon under the signature area so you have to sign all copies separately.
I'd assume a carbon-copied signature would not be considered binding at all, and would be also be dead easy to spot.
Re:Margaret Atwood (Score:5, Interesting)
Realistically, I suppose I'd be more likely to head down to the book store to see the weird robotic arm signing books than to talk with some random author I've never heard of.
Re:This is news? It isn't new. (Score:3, Interesting)
You're right. I remember seeing a long-distance handwriting machine at an airport 50 years ago, where someone in a remote city was writing messages to our city -- I think about the weather and flight delays. (I assume they could also have used teletype.)
And Harry Truman was the first president to use an Autopen to reply to constituent letters.
Re:The real question (Score:3, Interesting)
Who says they have to use stepper motors? There is a type of motor called a selsyn [wikipedia.org] or self-synchronizing motor. The way it works is this: You take two identical motors, called the transmitter and the receiver. You hook them up coil for coil. Then you supply power to the rotor coils. Any movement of the rotor on the transmitter motor generates a voltage in the stator coils (the stationary coils in the motor). These voltages are transmitted to the receiving motor, and produce a magnetic field that turns the transmitter motor's rotor by the exact same amount. This type of motor permits continuous angular displacement (ie no stepping). All you have to do is transmit the voltage levels long distance.
Granted, when you transmit things over the phone lines, there is an analog-digital-analog conversion that takes place. However the phone system samples at a high enough rate (8000hz) that a voice signal comes through. I think it could handle the 60Hz synchro motor signals with a high enough resolution that any digitization artifacts would be unnoticeable.
Re:Does not make much sense for authentication (Score:3, Interesting)
My understanding is that under US law, anything intended by the signer as a signature legally qualifies as a signature. That includes, but is not limited to standard signatures, electronic signatures, press seals, wax seals, visible fingerprints, etc. Now, this leaves open the question of weather a given mark is intended as a signature, and if so, what the signature is intended to mean. The signature may mean that I have seen and agreed to the contents, that I have seen the contents, that I am the author of the contents, or quite a few other things.
Now, the law will in some cases require specific types of signatures for some things, such as a true written signature, but not always.
For example for online trademark filing at the USPTO, the signature is any textual entry of the submitter's choice, as long as it begins and ends with the forward slash character. The USPTO considers that just as binding as a "normal" signature on a paper form.
In the same way, a PGP signature on a textual contract could be considered valid, subject to validity of the signature itself according to the OpenPGP standard.
For the record, IANAL.