3 ms·

Thanks. I see. I do also think though that the previous setup provided security by obscurity (it was hard to stalk, but it was also hard to notice being stalke

by pilif 17y ago

Thanks. I see. I do also think though that the previous setup provided security by obscurity (it was hard to stalk, but it was also hard to notice being stalked), whereas the new setup is open about that fact. Relying on your stalker not to be technically-sophisticated seems risky. I agree on the point of not changing existing tools, though arguably, this was a change for the better (now you see that you are being stalked and how to stop it), albeit one which could have done even better (by turning all of the existing features off, then implementing buzz and then giving people the chance to opt-in again).

maqr 17y ago
I'm also having a hard time understanding the issue here. Is any private data (besides social relationship data) being 'published' that wasn't available before via a simple Google search?
- mechanical_fish 17y ago
  As I will have cause to say again: Life is not a mathematics problem. There is a world of difference between that which is theoretically possible, that which is practical, that which is routine, and that which is obvious. And the tendency to assume that these things are all the same is a huge problem in tech product design, especially social software design. Cryptographers have a principle: security-through-obscurity is no security at all. And in the world of cryptography, that's a good principle. [1] But, alas, in the world of people who are being stalked, that principle is useless -- or, rather, it is isomorphic to in the long run you are doomed. Security-through-obscurity, and its undependable, amorphous cousins like social engineering, are all that you get, assuming that you physically exist, that you can't afford to live in an armed compound, and that you can't afford (or have no legal power) to get your stalker put in a secure prison for life. Don't force your users to give up all the pragmatic social tools that they know best, that they've developed for themselves -- their knowledge of the natures, personalities, and social norms of the people around them -- and make them live inside a math-class story problem. Such a move can literally kill them. And they won't appreciate it even if it doesn't. --- [1] Because cryptographers study a class of problems in which this assumption makes sense and maps well to the real-world situation.

by pilif 17y ago

maqr 17y ago

I'm also having a hard time understanding the issue here. Is any private data (besides social relationship data) being 'published' that wasn't available before via a simple Google search?

mechanical_fish 17y ago
As I will have cause to say again: Life is not a mathematics problem. There is a world of difference between that which is theoretically possible, that which is practical, that which is routine, and that which is obvious. And the tendency to assume that these things are all the same is a huge problem in tech product design, especially social software design. Cryptographers have a principle: security-through-obscurity is no security at all. And in the world of cryptography, that's a good principle. [1] But, alas, in the world of people who are being stalked, that principle is useless -- or, rather, it is isomorphic to in the long run you are doomed. Security-through-obscurity, and its undependable, amorphous cousins like social engineering, are all that you get, assuming that you physically exist, that you can't afford to live in an armed compound, and that you can't afford (or have no legal power) to get your stalker put in a secure prison for life. Don't force your users to give up all the pragmatic social tools that they know best, that they've developed for themselves -- their knowledge of the natures, personalities, and social norms of the people around them -- and make them live inside a math-class story problem. Such a move can literally kill them. And they won't appreciate it even if it doesn't. --- [1] Because cryptographers study a class of problems in which this assumption makes sense and maps well to the real-world situation.