The new HTTP_VIA field

I like the HTTP_VIA field sent to our WWW server from the cache.
Unfortunately it doesn't seem to add a whole lot of information, at least not
to someone as new to Web cacheing as me. The HTTP_VIA tells me the same name
as the REMOTE_HOST field (and the address of the cache can be found in
REMOTE_ADDR). I guess that REMOTE_ADDR is picked up from the socket or
something like that, and cannot be altered, but would it be possible to keep
the name of the original client machine in REMOTE_HOST? Then the Web server
would know from where the request originated, and could keep sensitive
documents within a limited audience if made non-cacheable. Actually I think
an X_REMOTE_HOST would be useful too, if the above method is technically
infeasible.

Best regards,
            Lars-Owe

-- 
-- 
 _________________________________________________________________________
/\                                                                        \
\_\ snailmail: Lars-Owe Ivarsson     | email: lars-owe.ivarsson@its.uu.se  \
  |  office: Uppsala univ., ITS      | phone: work:   018 18 78 70         |
  |   P.O. Box 887; 751 08  UPPSALA  |       home:     0224 930 55         | 
  |  home: Håcksby; 744 92 HEBY      |        int:  +46 224 930 55         | 
  |  for both: SWEDEN                |        fax:    018 18 77 25         |
  | _______________________________________________________________________|_
  \ \    (approximately 1300 kilometers northeast of Brussels)               \
   \/________________________________________________________________________/