To this point I would say, we need to know the physical location of the
machine (node) on the Internet sending the email.
do we really care where the message was sent from, as opposed to who
sent it?
I could see it being useful in the case of mail that was sent from a
compromised machine (if we could locate the machine very precisely), or
in the case of mail that was sent from a laptop that accessed a random
802.11 network ("drive-by spamming"). But I have a hard time
understanding how DNS could help us get this information, especially in
the latter case.
Also, just as I don't want to expose the actual identity of every
message sender, I wouldn't want to expose the location of every message
sender. (I can see it now - send an e-mail critical of Dubya, and a
G-man knocks on your door within the hour...)