[Top] [All Lists]

Re: Second Draft of "Implications of MIME for MTAs"

1992-05-26 18:02:26
There were some comments to my proposed definition which caught
some sloppiness when I wandered into the realm of "intent" and
also observed that the recipient's intent may not be knowable to
the gateway.  

So, I propose the following modifications:

    } A "mail gateway" performs store-and-forward electronic mail transfers
    } between two environments that are using *different* electronic mail
    } format syntax or semantics.  Hence, transfering email to any of the
    } existing, national email services, or transfering to any X.400 service,
    } or even transfering into a non-MIME RFC822 environment is performed
}an environment which encodes it's content differently than that of
}the originator,
    } by a gateway, rather than a simple relay.
}A key assumption is that the gateway must have knowledge of the differences
}between the sending and receiving environments and its appropriate actions
}are limited to those conversions which it can safely make, while maximizing
}the utility of the message communication between the originator and the

While I don't have any objection to the general use of Stef's approach,
which discusses things in terms of needing to perform "layer violcations"
and, in fact, believe it is useful to get people to think about such things,
I was/am primarily seeking to develop a definition which is procedurally
and objectively trivial to apply.  Note, for example, that the proposed
definition does not attend to the underlying email transport protocol,
per se, but stresses retention or modification of message semantics.

Thoughts, anyone?