Notes taken during the presentation:

Rohit Khare, introduces the speaker. In the process, he surveys which OSes are represented in the audience:
Microsoft NT, 9x, …: most
MacOS: some
Linux: more than Mac
Palm OS: couple

How many own a web domain? Approx. half, some more than one.

Slide "Authorization":
Rohit Khare, answered the rhetorical questions of the speaker, "How do you punish a key," with, "No more multiplications for you [the key]!"

Slide " Three Namespace":
Remark buy Carl Hewitt: "In the CIA, even the names of names are a secret."

Slide "Key as Global ID":
Some interlacing smart (funny, mostly sarcastic) remarks by the audience on "public key as global ID".

Roy Fielding: How do you get a collision free hash of the public key?
Carl Ellison: We assume that it's a collision free hash.

Slide "Local Names":

SDSI — comparable to work of Einstein! Discard notion of global namespace, and realize that each one of us has our own namespace … and use that in interactions with other people…linking of my namespace to yours.

Slide "Fully-qualified Names":

Q: Does key refer to Fred or to Betty?
Carl Ellison: Key refers to my key.

Bob Morgan: It sounds like you're claiming that the string which has the key in it is human-readable?
Carl Ellison: No, only the local name is human friendly. I do not believe it is possible to have a globally unique user friendly name.

Larry Masinter: How many keys are there?
Rohit Khare: [A key is encoded in] 34 bits.

Larry Masinter: What's the population of the world? And how many bits does that amount to?
Rohit Khare: {missed his answer}

Q: If you try to name something using normal English, the entropy increases … you get more density with email addresses, but that means some emails names are not human friendly …

Rohit Khare: question about how to expand what it means to interpret an SDSI name chain… Karl’s Jim’s Therese … <key>n1 n2 .. nN = kN

Esther Dyson: As you say if there is only one [root], it is irrelevant.
Carl Ellison: There will be more than one.

Post-presentation questions and answers:

Bob Morgan: How about identity changes?
Carl Ellison: You can't analyze a person’s name, your names remains your name ... There is this wonderful belief that a name cannot be broken. Every year, you can attach a different key to your name …but the trouble is it requires X.500 or a single certifying authority to have succeeded. There are millions of certificate authorities, and you tell them apart by their root key. So a name that lives forever impossible only if there is NO root key …. No need for key in (name <key>n1n2 … nN) = kN the key here has the same cryptographic lifetime, and the same vulnerability …

{Ed: Broader discussion of the issue. Speaker flips back to slide "Fully-qualified Names" to explain semantics of used syntax in more detail.}

Q: So I lose my smart card .. do I lose access to everything?

Bob Morgan: Your long lived name can and should be a public key  … same procedure that a CA would use, only he does it frequently?
Carl Ellison: Same procedure.

Rohit Khare: A multifunctional smart card is a suspect idea.  You don’t want to have a driver’s license for a credit card to be issued.