3 ms·

My point is much more narrow: that the collection of millions of records to make a graph of relationships, to scan for patterns in the graph is fundamentally a

My point is much more narrow: that the collection of millions of records to make a graph of relationships, to scan for patterns in the graph is fundamentally a different kind of search from obtaining the records of activity a lone person to specifically look in to their actions. My edit elaborated on this: the mass record gathering is not fundamentally millions of lone investigations happening in parallel, but a fundamentally different kind of investigation; they're collecting the edges of the graph, not merely many nodes instead of a single node. The court failed to rule on the relevant question, through what seems like ignorance of the underlying mathematics/information theory involved in the situation. Essentially, the government ruled on "What if you collect all the nodes for the graph, instead of just one node?" rather than "What is the proper manner (if any) to collect the edge information of the graph?" It's not obvious that the government has an authority to gather those relationships (the connections between specific nodes) in bulk, and that it is a different kind of search, fundamentally, from looking at a specific node (or many of them in parallel). What the information is going to be used for is relevant to the question of their obtaining it. In that sense, their ruling is bad because it's not even answering the right question.

neilk 13y ago
I hear you - of course it is different, but what legal principle can we call upon to make this count in the court? (I'm not a lawyer, so the following analogy may be way off.) Look at it this way. In the USA, people have a right to individual liberty. A cop can ask you to step out of the way while she performs her duties. That's a tiny breach of your liberty, weighed against the benefits of the cop performing her job of protecting the public. But you can't extend that principle to mean a cop can ask a million people to move across town, because a million times a small injury is a big injury. The government here is arguing that the injury of looking over one person's business records is zero. And arguably, in their terms, it is. The subject doesn't even know it happened. As long as the government treats the individual records with appropriate care, they would say there's no harm. And a million times zero, in their calculation, is still zero. What you're trying to say is that of course a million people's phone records is totally different from one. It gives the government a kind of power over the entire population they did not have before. But it seems we can't get there by multiplying the harms to one person. So what is the principle that we use?
- DerpDerpDerp 13y ago
  > what legal principle can we call upon to make this count in the court? It's already the case that different kinds of searches have different standards to obtain the information. The argument lies in convincing them of the truth: that this is a qualitative change in the type of search, not a merely quantitative change, ie, that they're not merely multiplying the number of a certain kind of search by a million (which could be the correct ruling, were the government using the information differently, as the court appears to believe they are), but rather, committing a fundamentally different type of search. > What you're trying to say is that of course a million people's phone records is totally different from one Not quite, it's a little more subtle: using a million people's phone records to correlate data and construct a graph not available without processing many of their records together is fundamentally a different activity than merely having a million records and looking through them each as if they were merely individual records. The NSA is analyzing the edges of the graph, and their relationships, not the node data directly; the court ruled that analyzing a million nodes is no different than analyzing one. I agree - however, it's not a relevant question to this case, and betrays a fundamental misunderstanding of what's happening on their part.

neilk 13y ago

I hear you - of course it is different, but what legal principle can we call upon to make this count in the court? (I'm not a lawyer, so the following analogy may be way off.) Look at it this way. In the USA, people have a right to individual liberty. A cop can ask you to step out of the way while she performs her duties. That's a tiny breach of your liberty, weighed against the benefits of the cop performing her job of protecting the public. But you can't extend that principle to mean a cop can ask a million people to move across town, because a million times a small injury is a big injury. The government here is arguing that the injury of looking over one person's business records is zero. And arguably, in their terms, it is. The subject doesn't even know it happened. As long as the government treats the individual records with appropriate care, they would say there's no harm. And a million times zero, in their calculation, is still zero. What you're trying to say is that of course a million people's phone records is totally different from one. It gives the government a kind of power over the entire population they did not have before. But it seems we can't get there by multiplying the harms to one person. So what is the principle that we use?

DerpDerpDerp 13y ago
> what legal principle can we call upon to make this count in the court? It's already the case that different kinds of searches have different standards to obtain the information. The argument lies in convincing them of the truth: that this is a qualitative change in the type of search, not a merely quantitative change, ie, that they're not merely multiplying the number of a certain kind of search by a million (which could be the correct ruling, were the government using the information differently, as the court appears to believe they are), but rather, committing a fundamentally different type of search. > What you're trying to say is that of course a million people's phone records is totally different from one Not quite, it's a little more subtle: using a million people's phone records to correlate data and construct a graph not available without processing many of their records together is fundamentally a different activity than merely having a million records and looking through them each as if they were merely individual records. The NSA is analyzing the edges of the graph, and their relationships, not the node data directly; the court ruled that analyzing a million nodes is no different than analyzing one. I agree - however, it's not a relevant question to this case, and betrays a fundamental misunderstanding of what's happening on their part.