At 9:55 PM 11/7/1996, Timothy C. May wrote:
"Truly noisy sources" are not at all unusual. Actually, the hard part is ever proving a source is _not_ noisy. (There are deep issues involving randomness here, and I usually go into the work of Kolmogorov, Chaitin, and others at this point. Consult the archives, or see a book on information theory.)
Truly, I am out of my depth. But, I will plunge ahead nonetheless. I think it is hard to show that a source is noise. Using only the signal itself, I believe it is impossible. That means there may be patterns which you cannot prove do not exist. That is surely less than ideal, especially if you are betting your life on it. And there are all sorts of ways you can blow your cover. Maybe your traffic patterns are sort of odd. Maybe you are up at odd hours. The weakness is stego does not have to be great before you are put on the suspect list. If cryptoanarchy is unpopular, you are in big trouble because the government can afford to watch you intensively.
The legal issue is this: can we pass laws and have them upheld by the courts which impose severe penalties on people for the supposed crime of having in their possession sequences of numbers which cannot be converted to meaningful English sentences? I maintain that the Constitution says we cannot. Of course, if the Constitution is thrown out, then the old Cypherpunk joke may come into play: "Use a random number, go to jail." (An Eric Hughes quote, from 1992-3.)
What Tim May maintains may be more reasonable that what the Supreme Court maintains. But it's the Supreme Court that will rule. We have laws which state, in essence, that certain large numbers cannot be copied legally if they are related to copyrighted executables.
In the model I am positing, there would be broad popular support for such policies.
I think you are assuming a lot.
I think I am, too. I don't subscribe to the Four Horsemen scenario. Peter Hendrickson ph@netcom.com