Re: [Emailcore] Ticket #8: Need a registry of header fields that are Ok to add after submission

[John C Klensin <[e-mail]>]

--On Sunday, 25 July, 2021 21:25 -0400 John Levine
<[e-mail]> wrote:

> It appears that John C Klensin  <[e-mail]> said:
>> Or you could convince the WG to tell me to go ahead and
>> create a normative reference to 5598 5598 to and use the
>> exception procedure, but I imagine some of us would argue
>> that would be tantamount to moving standards track and hence
>> out of charter.
> 
> Yikes. If we make this list of headers, it's not going to be
> normative, since we can't retroacively clean up 40 years of
> practice, and while 5598 is a useful model of the way mail
> systems work, there's a lot of places where theory and
> practice don't exactly match.\

Exactly.

> Many headers are ideally created by MUAs but in practice added
> by MSAs, so they get listed both places. There's also overlap
> between MTA and MDA depending on how a mail system is built.
> 
> I still don't understand what problem this exercise would be
> solving.

Speaking personally only, agreed.  For me that created a
situation in which IANA should insist on my precise definitions
to run even a piece of a registry and getting to those
definitions --with theory and practice matching-- is likely to
be sufficient trouble that the benefits and problem being solved
had best be _very_ clear.

best,
   john