Re: Last WG call for draft-ietf-dnsop-root-opreq-02.txt.

[Kent Crispin <kent@songbird.com>]

On Fri, Dec 03, 1999 at 11:24:03AM +0900, Masataka Ohta wrote:
> Kent;
> 
> > > > You are saying that telephone and fax were more serure than unsigned
> > > > email.
> > > > 
> > > 
> > > Telephone is secure only in it's billing, and fax is only as secure
> > > as telephony. Caller-ID and other signal messaging functions are
> > > also not guaranteed.
> > > 
> > > Anyone could call or fax and say they are someone that they 
> > > are not.
> > > 
> > > I think this is a little vague for a security measure.
> > 
> > Yes, it is vague, but that's the way it is.  "Security" isn't exactly
> > the characteristic being looked for in any case.  "Authenticity" gets
> > it a little better.  Obviously, a telephone conversation with someone
> > you know personally gives you *far* more assurance than an unsigned
> > email from that same person. 
> 
> Your wrong assumption is that operators always have personal relationship
> between them.

There is no such assumption implied by my words.  In fact I made no
claims whatsoever concerning whether operators know each other.  The
meaning of my statement was: IF the operators know each other, THEN a
telephone call gives better assurance. 

> With such an assumption, email conversation works almost as good as
> phone one (though it takes longer to exchange several mails).

With such an assumption unathenticated email is nowhere near as
assured -- email allows a man in the middle who can intercept and edit 
the messages; practically speaking, telephone does not.

Kent->MITM: "the md5 is X"

MITM->Masataka: "the md5 is Y"

Surrounding that bit of crucial information with all kinds of
personal information that is passed unobstructed won't help.  The
only thing that would help is if you actually used some form of
informal crypto.

-- 
Kent Crispin                               "Do good, and you'll be
kent@songbird.com                           lonesome." -- Mark Twain