[2445] in WWW Security List Archive
Re: cookies and privacy
daemon@ATHENA.MIT.EDU (Jacob Rose)
Wed Jul 17 21:16:31 1996
Date: Tue, 16 Jul 1996 13:28:40 -0400 (EDT)
From: Jacob Rose <jacob@whiteshell.com>
To: Dave Kristol <dmk@allegra.att.com>
Cc: www-security@ns2.rutgers.edu
In-Reply-To: <9607161658.AA12025@zp>
Errors-To: owner-www-security@ns2.rutgers.edu
> Let me illustrate.
> 1) You go to http://a.com.
> 2) The page you get back has in it <IMG SRC="http://b.com">.
> 3) Your browser makes a request to b.com for the image.
> 4) You get back, along with the image, a Set-Cookie header.
> 5) You go to http://c.com.
> 6) The page that comes back has another <IMG SRC="http://b.com"> in it.
> 7) Your browser requests the image from b.com and sends along the
> cookie you got in (4).
This is the step I'm talking about; the user's "apparent site" is c.com,
so in my scheme cookies would only be sent to c.com. That way, what the
user sees and what the browser does are directly related, and furthermore,
b.com is unable to correlate the hits from the same user visiting a.com
and c.com, since the cookie doesn't provide a key for doing so. b.com can
still find out that *someone* visited a.com and c.com, but not that these
were the same person, or worse, a particular person - and finding out who
you are is simply a matter of correlating enough data about you. Enter
your e-mail address at a single site that's cooperating with b.com, and
all the other records of your page hits and any input you may have made
there are linked right back to you.
I don't know if privacy is important, but if you feel it is, that's how to
keep it, I think. If you're running someone else's software (eg,
Netscape, Internet Explorer, Windows 95) on a computer on a network,
you are trusting the software to do only what you want with your personal
information. You never know what it could be saying about you to *anyone*
on planet Earth.
Jacob Rose "The truth is where the sculptor's
jacob@whiteshell.com chisel chipped away the lie."
