[19452] in cryptography@c2.net mail archive

home help back first fref pref prev next nref lref last post

Re: ADMIN: end of latest SSL discussion

daemon@ATHENA.MIT.EDU (James A. Donald)
Wed Dec 28 12:01:16 2005

X-Original-To: cryptography@metzdowd.com
X-Original-To: cryptography@metzdowd.com
From: "James A. Donald" <jamesd@echeque.com>
To: cryptography@metzdowd.com,
	"Perry E. Metzger" <perry@piermont.com>
Date: Tue, 27 Dec 2005 23:18:08 -0800
In-reply-to: <87y8263wmu.fsf@snark.piermont.com>

    --
In the SSL thread various solutions were proposed, or 
rather existing solutions pointed to:

1.  SSH just works.   So generalizing from the success 
of SSH, the browser should remember who you have 
relationships with, and the keys of the people you have 
relationships with.   To avoid the name overload 
problem, those relationships should be named by Zooko's 
triangle, as the petname tool does, and should be a 
special kind of favorite, as the petname tool makes 
them.   This requires that establishing a relationship, 
and verifying a shared secret, should be part of the 
browser chrome, as it is with SSH, rather than a 
particular application of generic web forms, as it is 
with existing practice.   So when you hit a phisher, 
significantly different chrome comes up.

2. Phishers are after shared secrets, so secure each 
shared secret, and thus each relationship, with 
SRP-TLS-OpenSSL  This also requires that establishing a 
relationship, and verifying a shared secret, should be 
part of the browser chrome, rather than a particular 
application of generic web forms. 

    --digsig
         James A. Donald
     6YeGpsZR+nOTh/cGwvITnSR3TdzclVpR0+pr3YYQdkG
     8epIQqxZ+sfUW+5ao0hWd4g/hAhRlqifZr6xWoQn
     47kvMBcL6UqQ54XSgEcxbJd8xqAh2LSxufi/3IBdG



---------------------------------------------------------------------
The Cryptography Mailing List
Unsubscribe by sending "unsubscribe cryptography" to majordomo@metzdowd.com

home help back first fref pref prev next nref lref last post