Re: [perpass] Draft charter for a Transparency Working Group

[Stephen Kent <[e-mail]>]

Ben,

>> The text describing how 6962 uses Merkle trees is good. I think the
>> phrase "prove its own correctness" is way too broad. The example
>> you cite shows how to demonstrate internal consistency for a log,
>> and to enable third parties to verify certain lob properties. That
>> is much narrower than what the term "correctness" implies.
> How about, instead of "can prove its own correctness
> cryptographically", we say "allows efficient verification of
> behaviour"?
>
I still find that phase vague. What sort of behavior is being
verified? Isn't the behavior amenable to verification a function
of the context details? For example, a self-signed cert is
an example of a crypto construct that allows an RP to verify a few
aspect of its "behavior"
      - the public key contained within the cert is matched to the
        private key used to sign it.
     - the cert content was not modified after it was signed

But most of the other semantics of CA-issued certs are not
verified by this construct.

Steve