[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-devel
Subject:    Re: ssl auth failure gui: does "continue" do what I think it does?
From:       Matthew Woehlke <mw_triad () users ! sourceforge ! net>
Date:       2009-06-09 16:47:04
Message-ID: h0m3m9$v4$1 () ger ! gmane ! org
[Download RAW message or body]

Jeff Mitchell wrote:
> Matthew Woehlke wrote:
>>> But there isn't a choice.  Certificates are essentially the only
>>> encryption method feasible for most sites, because of e.g. browser
>>> support.  So if all you need is encryption, and not authentication, you
>>> still have to use the same system.
>> But *you don't get encryption* this way
> 
> But you do. [S]aying it's not [...] is factually wrong.

Humph. Apparently wiktionary agrees with you, defining "encryption" as 
"the process of obscuring information to make it unreadable without 
special knowledge, key files, and/or passwords."

...which means that rot13 is "encryption".

So... let me clarify. You don't have* /end-to-end/ encryption. There is 
no way to know that the MitM isn't passing your bits along in the clear. 
You only know you have encryption in the strict technical (and useless, 
as far as security) sense.

(* To be clear, you /might/ have end-to-end encryption, but you have no 
way of knowing.)

> And sometimes you don't care.  This was my point, which you've basically
> ignored.

No I haven't. If you re-read what I wrote, I said it should be made 
clear that unauthenticated encryption is useless and should only be 
permitted when you don't care if the connection is secure or not.

> (Secondary question: Who is going to ensure that the dialog box
> communicates the uncertainties of self-signed certificates without
> making users feel like the world might end if they decide to proceed?)

"Unless you have some other way to verify the authenticity of this 
certificate, accepting an unsigned certificate should be considered 
equivalent to using an insecure connection." (And maybe something about 
"unauthenticated HTTPS == HTTP".)

Maybe we should instead just present such sites as insecure? Maybe even 
without giving warning?

-- 
Matthew
Please do not quote my e-mail address unobfuscated in message bodies.
-- 
Sorry, but I can't look into that right now. I'm running low on 
sacrificial chickens.

 
>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<
[prev in list] [next in list] [prev in thread] [next in thread]