From: abostick@netcom.com (Alan Bostick) Yes, but you are denying the way in which delaying, like bouncing, actively interferes with the timely forwarding of non-signers' messages, while merely marking them is a more passive form of harrassment. A delay for one minute (assuming notice for the delay) is hardly different than notification only. A delay for a month is hardly different than a bounce. Not all delays are the same. They cannot be analyzed as a single category but are better analyzed with respect to the characteristic time scales of the discussion. You keep insisting that delaying unsigned messages does not interfere with non-signers' abilities to participate in the discussion. I say you are wrong. It's a positive hindrance. This is statement is true for large delays and false for small ones. The interesting issue to me is where a boundary might lie. (Are you going to make sure that all the signatures are valid, or will you accept someone sticking a PGP signature into their .sig and using it over and over?) At first, it would just be a recognizer for syntax, but at both ends. A second effort might actually hash the message but not bother with the signature itself. The second effort would require almost all the processing involved in a real signature and require the same architecture. It would not, however, be subject to the key distribution problem that I don't want to make a prerequisite. It occurs to me that a format with just a hash might be generally useful against random data corruption, and not just a workaround hack. Eric