Using E-mail Headers to Identify Web Browsers

from an e-mail i sent to my brother

To      : troq
Subject : more data usage

Mail headers have a variety of IP addresses in them, including (depending
on the method of mail used) the ip used by the sender's computer (as
opposed to cimply the server the mail came through).  If a person always  
uses the same computer (or, rather, if there is a one to one relationship
between ip and individual) then this information can be used to track an
individual...

...in web logs.  So i parsed through all the mail in my inbox to correlate
ip's to email addresses and full names, and compared this to the 900+
unique hosts in web logs this month, in order to see if any of these were
the same.

i got back 12 results.  that is, 12 ip's that have sent me mail have also
browsed my website this month.  ot of those 12, 11 had a one-to-one
relationship (one was orchard's ip, from which multiple people have sentme
mail).

what does this mean?  this means that for 11 unique visits to my site,i
can say with a fair amount of certainty that i know who was looking at my
website.  note that out of those 11, 8 are friends of mine (and mom), but
the other 3 are people i do not know but are active posters to lists i am
on.

this technique is only useful in identifying people who have sent me mail
(that is, friends, relatives, and people on the same mailing lists as me).
also, there is no guarantee that the person's ip hasn't changed or that
more than one person don't use the same ip.  still, nothing *special*
needs to be submitted by the viewer - no javascript corruption, no buffer
overflows - and i can still say with a fair degree of certainty who the
viewer is.
  
at least 1% of the time, i can.



notes:
i only used my inbox to tabulate the mail From:'s and ip's.  if i
continually did this for all the mail i ever get/have gotten, i could
create a much better database of potential ips <-> people, increasing the
1% rate to perhaps as high as 4% (based upon how much saved mail i have
that could be parsed).

If i could, 4% of the time, deliver a message like "Hi John Doe, how are 
you?"  without ever having had John Doe directly tell me he was browsing
from that particular computer, then i would be happy.

-f
http://www.blackant.net/