<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<html>
  <head>
    <META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=ISO-8859-1" />
    <title>diffie-hellyes</title>
  </head>
  <body>
    I haven't read the OTR spec in awhile, but I seem to recall that one digitally signs the AES key derived from the Diffie-Hellman transaction.  Why not simply sign only your own public value in the Diffie-Hellman process?  If the other party sign's his, then you know with confidence the shared secret (private key) that you will both generate.  This seems to provide substantially better deniability, because there's no way to prove you ever even knew the other party's public value and generated the shared secret.  You could deny that you had ever even seen that private key.  Does this make sense??  Am I missing something?  Is this how it's already done and I just misunderstood?<br /><br />Just curious..<br /><br />I think the next step on this plug-in, and I wish I had the time to help, would be to make the equivalent of mixminion for IM.  Essentially, your message is encrypted like those little Russian dolls, so that the next person in the line can encrypt a layer, and then on down the line, until the final party can view the message, and an outside observer would be hard pressed to discover who you were talking to.  Maybe this is too process intensive, since it would require successive RSA decryption to peel away the layers, as well as the permission of various people on a network.  Just a thought..<br /><br />Clay<br />
  <pre>--------------------------------------------------
ENCRYPTED MESSAGES ARE PREFERRED.  PLEASE USE
THE PGP PUBLIC KEY FOR BROKEN LADDER AT
HTTP://ESKILO.WARPMAIL.NET/
</pre>
</body>
</html>