Even my local phone company customer "service" has software that can recognize a list of words in english and spanish. None of us knows what the NSA can really do, but they have a lot smart people working on... well, something. They have been gathering huge amounts of information for decades, and we can assume that they do as much as possible with it. I think they define what they are looking for precisely and get a very small relative amount for human analysis. Spying on global communications is universal and entirely accepted. The scary part is the potential for misuse of the information. And by that, I mean something like targeting you political opponent. When international and domestic communication cannot be separated easily, such things are possible.

Indeed I am, which brings up such questions as, whether computers doing word searches is spying. Is doing connection analysis of random numbers, spying?

Well, you have lost me now. Truly random numbers are not connected. An ideally encoded digital channel appears entirely random, but yields its information patterns when decoded properly. Is this what you mean?

Of course this is part of spying, and it is illegal when applied to certain kinds of information.

Sorry, I sometimes forget to elaborate on conceptual shortcuts. My image of data-mining telephone networks incorporates the idea that telephone numbers are issued randomly. So, comparing how many calls one phone number makes to another number is an anonymous process. It would seem that an anonymous process like that would be outside the fourth amendment's scope.

The information required to make those correlations must be saved for billing purposes. It does not require accessing the real time data stream. Obviously much more than that is on the table: the actual data.

A list of anonymous "who calls who" is essentially useless. You need to know who a particular individual calls, and this certainly is private if the individual is a US citizen.

Coming so late in the Bush admin., could this be primarily a defensive distraction intended to assist the guilty parties safely out of office? If so, the best response might be a strong a strong offense.

The only thing NSA can be doing is link analysis. Who is communicating with whom.

With access to the actual data stream, much more is possible. It is purely a practical problem in information analysis. You do as much as you can, and that is a lot more than you could do in the past. Link analysis first become possible how many computer generations ago?

Sum up the six states and you see the republican turnout is not much changed from the previous records, but even without adding it up, and it is clear that the dems are up significantly.

Especially when you realize that the link analysis does not require the "fat pipe" at all since the telecoms must save that info separately for billing purposes. So why pay the big bucks for everything if you could get what you need more cheaply off a telecom administration server? (Shooter is overpaid for his services, no matter how small his salary.)

For a "known", there is no problem obtaining a warrant for monitoring communication within the US. Outside the US requires no warrant, of course. The FISA "kerfluffle" is primarily about extending the "no warrant" case to fiber optic system controllers located in the US that carry much foreign-to-foreign and foreign-to-domestic calls as well as purely domestic. The problem is that there is no known way to really prevent the domestic from being collected if the other is. It is not really that simple, of course. But that is pretty much the problem.

Warrants require specific targets. I have no legal expertise, but my layman view is that the FISA court can and does allow just that kind of monitoring. Perhaps one of the lawyers here could answer that more accurately.

But there is at least one more important step. Once there are enough destroyed countries under control, the imperial power pays their people a minimal amount to fight the next war. No need to hide the body bags because they go elsewhere, and also the cost decreases.

As digital photos improve, and we have the bandwidth to transmit higher-quality recordings, who's to say that the digital trade-offs we now accept won't go the way of Wonder bread as well.

Please excuse a bit of criticism, you two can express ideas orders of magnitude better than I can, but there is a bit of the luddite in both of you, in different ways. Digital video may have been pushed into the market a bit early, but it is certainly vastly superior to film in its potential and even in practice now in some ways. The digital revolution in audio was about 25 years ago, and the instant superiority of CDs was apparent, but we still have folks who worship scratchy plastic records, and highly (but pleasantly) colored tube amplifiers. (OK, guitar amps need the coloration, but not stereo HiFI, and the compression in audio recordings for transmission is unacceptable, but this is, as WT says, temporary.) How much of the superiority of film is just in being keyed to its particular limitations and strengths?

My apologies, I did not think either of you would react so strongly. But I enjoyed your replies, so thank you both.

Ondelette, I am a bit puzzled by the beat frequencies. Are you saying that it is necessary to have the frequencies above, say 20 KHz, in the playback system because these tones can beat in the ear and thus non-linearly generate lower frequency tones that are audible? Do you have a reference?

Powerpoint really is dreadful. But I like Illustrator, and dragging AI files into Apple's Keynote is easy. For equations, Latex output can be rasterized and used in either; slow but effective.

But powerpoint is intended for about 32 bits of information per slide. Apparently that is the proper amount of information for the business community.