[e2e] disintermediation?

[Fred Baker]

At 04:08 PM 10/12/2001, Bob Hinden wrote:
>I wonder if a NAT would constitute "publication" by this definition?  The 
>IP payload is being modified.

At 10:22 AM 10/13/2001, Bob Hinden wrote:
>NAT modifies more than the the IP header.  In many cases the port numbers 
>and checksums in the transport protocol is changed.  In other cases (e.g., 
>FTP) the contents of transport payload is also changed.  This change may 
>result in changing the amount of data in the payload requiring the TCP 
>sequence numbers to also be modified.

I should think "publishing" would refer to the data the user interacts 
with, not the plumbing that moves it around. It might apply, for example, 
to Lynne's ISP that identifies advertisements in HTTP payload and replaces 
them with other advertisements, but not to a QoS box that changes TCP 
window advertisements to pace TCP transmissions, or a NAT which twiddles 
the IP header. FTP control that changes the numbers with a view to offering 
exactly the requested service, but does not change the content being passed 
around, is being a stateful gateway to move the content, not a publisher of 
that content.