If you're already adept with text analysis you should be able to do the complementary operation. Make a system that gathers and analyzes gigabytes of real-world text in order to normalize input text for use with a nym and broaden the correlation with public writings. As it comes more into use it will start analyzing itself with the end result that vocabulary and variety will be minimized. The joys of positive feedback. Call it baa-sheeplifier-baa.pl. It may not make reading a pleasure but it could help in the neverending fight against *evil*. Mike
(Re: CDR RE: snipped from headers.....)
On Wed, 18 Apr 2001, Sunder wrote:
Ray Dillinger wrote:
On Sun, 15 Apr 2001, Declan McCullagh wrote:
And your possible motive for spreading the word about his reputation, which ties you to an illicit transaction, is what exactly?
Wouldn't your own reputation be blinded by a nym anyway?
Give me a few dozen writing samples from each of a hundred known people, and another writing sample a hundred words long from one of them under a pseudonym, and I can tell you to a 90% probability which of the hundred known people wrote it.
If some persistent pseudonym has a record with hundreds or thousands of illicit transactions, the lions are going to be crawling cyberspace for *any* writing that matches its style closely enough to have been written by the same person. They'll get a short list. Then they'll start eliminating possibilities and when they're down to three or four they'll start getting wiretap orders.
With the wiretap order, they can run a sting or a man-in-the-middle attack so they've got one solid charge. That will net them an arrest warrant if it works. But whether this works or not, they can still get a search warrant after they give it a shot.
If the machine is not theft-secure (and face it, almost no machines are), the arrest warrant issues anyway and the owner of the pseudonym winds up in jail. And the lions didn't have to do any particularly clever cryptanalysis to get there. All they had to do was run a spreadsheet counting grammar, word choice, sentence length, and a few other parameters until they found a match.
Bear