Eric, count this as another vote for default encryption of all communications links. Omega: the way to tell is to run a frequency count on the text. If it follows the usual distribution for the language it's in, then it's probably plaintext. In which case the BBS rejects it. Voice: yeah, it's a pain in the tail. One thing I thought might be interesting is to use two digitisers: one for the voice input, another for a keystream which is derived from a radio or TV program signal which can be picked up simultaneously by both correspondents. XOR the two streams together and then do whatever you have to do to make the encrypted results transmissable. I actually tried building an analog version of this about ten years ago (might even bring it to one of our meetings, just for fun). Analog voice "encryption" is actually pretty worthless (I didn't know how worthless until I experimented with it) from a security standpoint. Voiceprint modification may have some uses. About five or six years ago I built one of those: based on a pitch shifter with a graphic EQ on the input side and another on the output side. Worked, sort of. Now there is a phone you can buy with something similar built in. -gg