On 09/24/2018 06:49 AM, Ralph Seichter wrote:
On 24.09.18 02:12, Dave Warren wrote:
I don't see anything obvious that addresses my approach (only the approach of sending a message from a consistent address out slowly, which has several obvious flaws).
Messages are already uniquely identifiable, and your approach is just a variation of the method Andreas described. While it bundles spamtraps, it is still just as easily avoided using trigger address sets in the manner I mentioned before.
-Ralph
Maybe I misunderstood the proposal. Or unconsciously embellished it.
I was thinking that there'd be a set of Tor Project honeypot accounts, with the same apparent account (e.g., Jay Baker). But in fact, there would be a distinctly identifiable "hidden key" for each subscriber of each list. Periodically, the set of honeypot accounts would send innocuous messages to the Tor lists.
So let's say that Jay Baker instance with hidden key "Aj0qAU3Dc7PJzK" had sent a list message to just one subscriber. And then it received sex spam. That would arguably implicate that subscriber in the spamming operation. No? And then that subscriber would be unsubscribed.
Of course, any sane spammer would use throwaway accounts. And they'd just replace them as needed. However, once the system were operating, new subscriptions could be correlated with subscription removals. Perhaps subscription removals could be done in batches, to make that more obvious.
But of course, that would be just too creepy.