ietf-822
[Top] [All Lists]

Re: General considerations for new message field specifications

2004-11-02 07:16:57

Can anybody think of additional considerations (other than

the transparency issue: what happens, if one/some/all of the MUAs, MTAs
involved in generating, transporting, delivering or digesting the message
do *not* understand this header field?

agreed. in general, only the sending and recipient MUA should have to understand the new field; MTAs shouldn't have to care about it. if you're defining new functionality for MTAs it generally doesn't belong in the message header. but in reality there are some exceptional cases. for instance MTAs do rewrite addresses (even though they shouldn't) so this has implications for any new field that has addresses. and with the to-noreply and cc-noreply proposal it became obvious that the submission server would also have to understand them (at least for sendmail -t scenarios).

so basically any new proposal for header fields should analyze its impact on sending MUA, submission servers, intermediate MTAs, final delivery agents, and recipient MUA (including how this affects replies if applicable). when applicable, it should also analyze the impact on mailing lists, forwarding, mail filters, and other intermediaries, and auto responders.


<Prev in Thread] Current Thread [Next in Thread>