Re: homework, not an experiment, draft-crocker-email-deliveredto

On further thought, this experiment makes no sense.

We have two decades of experience with Delivered-To. It has multiple
compatible implementations that have have worked the same way for decades.

Since there is no chance that the implementations will change,
assuming we still care about interoperation, we can write a draft
describing how it actually works, based on the existing
implementations. That way, if other MTAs want to implement it, they
can interoperate. The obvious candidates to write it are people
familiar with the existing defacto spec, such as Viktor, Sam, and me.

If someone wants to run an experiment with some other delivery marker
with yet to be defined semantics, sure go ahead, but call it something
else like Delivery-Address to avoid confusion with the existing
well-defined Delivered-To.

R's,
John

PS: Isn't this how IETF standards used to work, write a couple of separate implementations,
see if they interoperate, and then write it up as a draft standard?