From: ichudov@algebra.com (Igor Chudov @ home)
To: frantz@netcom.com (William S. Frantz)
Message Hash: b71758835ac12b908ef08e722ad445fe0520a013875bc0b3716d120445ec25fa
Message ID: <199704160255.VAA00973@manifold.algebra.com>
Reply To: <199704160244.TAA10775@netcom6.netcom.com>
UTC Datetime: 1997-04-16 03:06:42 UTC
Raw Date: Tue, 15 Apr 1997 20:06:42 -0700 (PDT)
From: ichudov@algebra.com (Igor Chudov @ home)
Date: Tue, 15 Apr 1997 20:06:42 -0700 (PDT)
To: frantz@netcom.com (William S. Frantz)
Subject: Re: Introducing newbies to encryption (was: Re: anonymous credit)
In-Reply-To: <199704160244.TAA10775@netcom6.netcom.com>
Message-ID: <199704160255.VAA00973@manifold.algebra.com>
MIME-Version: 1.0
Content-Type: text
William S. Frantz wrote:
> > It's absolutely true that nothing on a centralized Unix machine is truly
> > secure. However, is abandoning all pretenses of crypto and security in
> > favor of holding out for a utopian ideal really the best solution? Does
> > using encryption for email on multiuser machines actually hurt the cause
> > of the security community in the long run?
> >
> > (I'm not asking rhetorical questions here -- I'm truly looking for some
> > thoughts on this.)
>
> Since security is not binary (i.e. talking of secure and insecure is
> nonsense. You must talk of more or less secure.), you have to look at the
> threats. If you are sending email from a multi-user Unix machine, encrypting
> it removes some threats (e.g. wiretaping) without adding any new threats.
> (There are still the continuing parade of UNIX holes based on the C string
> model.)
>
> I would say that if users don't think they are safe, just think they are
> a bit safer, then encrypting on a multi-user machine is a good thing because
> it is more secure than not encrypting. It is still less secure than a
> single-user system with Tempest shielding.
right, the real problem is that users start thinking that they are
really safe, and start doing dumb things.
- Igor.
Return to April 1997
Return to “Toto <toto@sk.sympatico.ca>”