4 ms·

I think, fundamentally, physical locks _can't_ be solved. At the end of the day, given enough time, there will always be a physical workaround to a physical lo

I think, fundamentally, physical locks _can't_ be solved. At the end of the day, given enough time, there will always be a physical workaround to a physical lock. Even with a "perfect" electronic lock, somebody could always just cut through the door. IMO the purpose of a lock is to either 1) slow down an attacker long enough for (eg) police to show up, or 2) discourage an attacker from breaking into _your_ house in favor of your neighbor's house.

snet0 5y ago
I think mentioning "well if the lock's too hard they'll just break the window" is missing a point: locks right now are just too easy. Destructive bypasses to locks, or even just cutting the lock into pieces, are probably a good thing, since it prevents the equivalent of losing your Bitcoin wallet. But you want the attacker to be forced to either make a loud mess or go elsewhere. If you didn't encrypt HTTP, you could sniff packets on the network, walk through the front door. If it was weakly encrypted, someone with some level of expertise could basically do the same, pick the cheap lock and walk through the front door. With RSA-like encryption, this avenue of attack is basically closed, and now the options are more like "try and steal this guy's laptop" or "launch MITM attacks".
Animats 5y ago
I think, fundamentally, physical locks _can't_ be solved. Sure they can. Lock-picking is a search for a minima. You just need a design where there's no way to tell if you're getting closer to the right combination. That is, something senses the combination on the key, saves it, and then tests the saved info while protected from manipulation. Doing this cheaply and in a small package is difficult. Someone recently built such a lock and sent it to the Lock Picking Lawyer.[1] No results yet. One classic lock close to that was the Chubb Detector Lock. If any lever was pushed too high, the relocking device tripped and the lock would no longer open. Use the wrong key and you were locked out.[1] This was usually fitted with a second mechanism so that turning the correct key in the wrong direction would reset the detector. If built without the reset feature, pick attempts or using the wrong key would disable the lock permanently. This was highly secure but inconvenient. This is probably a solveable problem if you're willing to have a sizable box on the the door, like 19th century door locks or jail locks today. You'd want that anyway, for mechanical strength. [1] https://hackaday.com/2020/11/29/making-a-unpickable-lock/ https://hackaday.com/2020/11/29/making-a-unpickable-lock/ [2] https://youtu.be/7Q6rsbeJZMs https://youtu.be/7Q6rsbeJZMs

snet0 5y ago

I think mentioning "well if the lock's too hard they'll just break the window" is missing a point: locks right now are just too easy. Destructive bypasses to locks, or even just cutting the lock into pieces, are probably a good thing, since it prevents the equivalent of losing your Bitcoin wallet. But you want the attacker to be forced to either make a loud mess or go elsewhere. If you didn't encrypt HTTP, you could sniff packets on the network, walk through the front door. If it was weakly encrypted, someone with some level of expertise could basically do the same, pick the cheap lock and walk through the front door. With RSA-like encryption, this avenue of attack is basically closed, and now the options are more like "try and steal this guy's laptop" or "launch MITM attacks".

Animats 5y ago

I think, fundamentally, physical locks _can't_ be solved. Sure they can. Lock-picking is a search for a minima. You just need a design where there's no way to tell if you're getting closer to the right combination. That is, something senses the combination on the key, saves it, and then tests the saved info while protected from manipulation. Doing this cheaply and in a small package is difficult. Someone recently built such a lock and sent it to the Lock Picking Lawyer.[1] No results yet. One classic lock close to that was the Chubb Detector Lock. If any lever was pushed too high, the relocking device tripped and the lock would no longer open. Use the wrong key and you were locked out.[1] This was usually fitted with a second mechanism so that turning the correct key in the wrong direction would reset the detector. If built without the reset feature, pick attempts or using the wrong key would disable the lock permanently. This was highly secure but inconvenient. This is probably a solveable problem if you're willing to have a sizable box on the the door, like 19th century door locks or jail locks today. You'd want that anyway, for mechanical strength. [1] https://hackaday.com/2020/11/29/making-a-unpickable-lock/ https://hackaday.com/2020/11/29/making-a-unpickable-lock/ [2] https://youtu.be/7Q6rsbeJZMs https://youtu.be/7Q6rsbeJZMs