Wow, Tim May agrees with me with respect to the PGP 'Web of Trust' *AND* even goes so far as to obliquely recommend the 'Small World Network' model of key managment.
Will wonders never cease...I hope not. ---------- Forwarded message ---------- Date: Sat, 16 Feb 2002 11:12:53 -0800 From: Tim May <[EMAIL PROTECTED]> Reply-To: [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: CDR: Re: On NOT publishing the public half of a key-pair On Saturday, February 16, 2002, at 09:17 AM, [EMAIL PROTECTED] wrote: > On 15 Feb 2002, at 22:42, Morlock Elloi wrote: > Allow me to clarify my point. I wasn't actually suggesting anyone > do anything in particular, all I was saying is that if you don't > publish your public key, and in particular if you use a separate > public key for each person with whom you correspond, you're > not really gaining any conceptual benefit from using public key > encryption. You still are. If Alice encrypts to Bob's public key, and then sends the encrypted message or leaves it on her disk, whatever, then the message is not readable by anyone who does not have Bob's _private_ key. By contrast, use of a old-fashioned secret key means that anyone who saw Bob giving the secret key to Alice (wiretappers, for example) now has access to all communications using that secret key. This was all covered in the mid-70s in the seminal Diffie and Hellman papers on how parties who have never met may nevertheless communicate securely without physically exchanging a key. (Interstellar communication being one of the examples...) There are interesting issues, to be sure, such as man-in-the-middle attacks (e.g., the Rigelians one things one is communicating with are actually Zeta Reticulans who inserted themselves into the link.). So public key systems have many elegant aspects that go well beyond the "universal phonebook" model. I agree with the earlier poster (forgotten who) who said that publishing public keys is not at all necessary, or even recommended. I can think of so many reasons why this is so that I didn't even think it was important to voice my support for his or her point. > Right, and that's great when it's convenient. But if for some reason I > want to send encryptd mail to someone n, say, finland, I'd > rather not have to travel to finland to swap keys with him if > I can avoid it. You'd rather swap secret keys? Or is your claim that having your name, call it "Blacknet," in a registry means something? Absent a whole infrastructure of biometrics, of certificate authorities, etc., key registries are flawed. The spoofing or MITM sorts of attacks are never easily solved, anyway. Traveling to Finland is hardly the efficient way to handle this. --Tim May "How we burned in the prison camps later thinking: What would things have been like if every security operative, when he went out at night to make an arrest, had been uncertain whether he would return alive?" --Alexander Solzhenitzyn, Gulag Archipelago
