Reply #7: I really don't think I overstated much.

[hootinholler]

as someone who works every day with vast sums of data I feel its technologically impossible to accomplish the level of eavesdropping you are inferring

As someone similarly employed, I don't think that I did overstate the current feasibility of this type of endeavor. Technologically speaking it becomes a question of efficiency. The business proposal would consider ROI, return on investment. There are numerous techniques to improve both of those considerations over brute-force methodology, behold Google.

Consider conceeding for a moment that the capacity to simply collect the traffic exists. I think that collection would in and of itself constitute a crime, and why I think the courts have a problem with the dragnet aspect. Perhaps some lawyers could chime in on that point?

Once the traffic was collected, the metadata alone would be a treasure trove. For those less technically challenged, metadata is data about data. In this case it would be who called whom, when, duration, etc. Then the distributions of that sort of data would be analized to identify the nodes in a network of individuals.

Once those communications were identified, they would be targeted for indexing. The targets would be selected by things like call distribution frequency, degrees of seperation, etc. Throw in some random traffic samples for good measure and now you only have to actually index a significantly smaller percentage of the raw data, I'm thinking on the order of 10%.

Plus as technology and methodology improve, the ability to retroactively process more selections increases. It also increases as more machines are brought into the processing network through acquisitions.

It's the retroactivitity component of the capability, coupled with the capacity of the neocons to order it, that I find scary.

-Hoot