This headline should presumably read smartwatch, not -phone?
Doesn’t this suppose that you type your PIN with one finger, without changing the shape of your hand and moving your hand uniformly? Interesting though…
Beat me to it Joe. The paper references information that the phone can gather too but I think Cory meant watch
Hmm, yeah. Tried a couple of keying motions with my hand and there is definitely a bodily displacement you can feel around your core, especially if you move from side to side and forward backward with your whole arm, rather than just moving your hand/fingers, which feels weird and unnatural.
SILENCE FOOL! BOING BOiNG IS NEVER WRONG!
Jesus Christ, this “article” is almost like saying that if you point your phone’s camera on the PIN pad while you type clearly with one finger and record it you can get the number. What’s the point here?
Also, combining different languages in “an algorithm” is not an achievement to brag about. I do this when I’m being lazy and found different parts of a solution written in different languages and can’t be bothered to rewrite it all in one language.
read: he found a bunch of free processing libraries in various langauges, and didn’t want to rewrite them.
not very enterprisey: only 3 languages and no frameworks mentioned
most right-handed people i see with watches (smart or otherwise) wear them on their non-dominant hand. I do. I enter PINs with my right, watch is on my left.
Both my wife and I do this. One reason for doing it is simply that actions like mousing can reduce step count accuracy in smartwatches, so it makes sense.
I think the standard has been for right-handed people to wear watches on the left hand for decades before the introduction of smart watches and step counters.
But regardless, I agree with you and @kvanh: this is most likely a non-issue for the majority of users who wear a smart watch on their non-dominant hand.
You think that’s creepy? I didn’t even know that I have a smartwatch!
It’s like dogs - they understand human languages but they pretend not to.
Your smartwatch can somewhat, sometimes, guess your ATM and phone PIN if you decide to make a goofy hacky app that does jury-rigged keylogging.
ftfy
According to my grandfather it was because the dominant hand gets used a lot more than the other and so the accumulating shocks reduce the accuracy of dominant-hand-worn watches. This doesn’t of course apply to non-mechanical watches but I have always heeded the advice of someone who expected his section of the railway to run to Swiss timekeeping standards (and if I was on a train journey with him expected me to know how fast the train was going by timing the quarter mile markers.)
So…good
BEGIN OT
Notice how software has reversed the meaning of framework; originally a framework was a simple physical structure intended to support something important in an efficient way. Now a framework is a complicated structure which forces you to think in a particular way while introducing massive inefficiencies and possibly serious vulnerabilities.
END OT
I’ve always been paranoid about somebody near me trying to figure out my PIN so I typically try to key in the number using only finger movement. This hack only works if the person makes deliberate movements with the entire arm. Now if they could hack the wrist HR sensor to be able see how much my finger tendons are moving, I’ll be screwed
But a right-handed person would typically wear their watch on their left wrist, no? I only see the watch “learning” my PIN if I wear it on the same wrist I use when typing on the ATM keypad…
I see that @kvanh got here first. 
*point the number of one ) in the olden times , fighter pilots ( and others ) flew with the throttle and prop controls handy to the port hand , and the ( control ) stick in the starboard ; with the watch face to the inside of the wrist , and often on the starboard hand ~
*point the number of two ) coding is kewlio !! demos are kelwio !! proofs of concepts are often not polished and ready to ship , but can none-the-less inspire and show extremely high levels of competent functionality in the triad of algorythm , data , and syntax ; using , or even being aware of , multiple libraries in multiple languages ( syntax ) for different parts of a system is completely cromulent and even laudable and kewlio ~
*point the number of thrreee ) fitness sensors are sometimes separate from and in addition to a watch and so can be on either arm ~
- point the number of four ) there is NO point four
( point the number of five has been folded into point 2 due to a noisey bluetooth connection )
