In the problem statement, I think it would be a good idea to address the existence of the "abuse-mailbox:" attribute as well. This is causing a lot of confusion over the proper usage of "abuse-c:". With the right implementation, we should end up in a position to remove "abuse-mailbox:" altogether.
The problem imho with the abuse-mailbox is, that it still exists in places where it should not exist and is misused in certain cases. The difference between an email attribute and the abuse-mailbox attribute is, that email is for person to person messages, while abuse-mailbox is used for automated reports. This is been used very actively and very successful by a lot of network operators and organizations that send automated reports. I'd object to remove the abuse-mailbox attribute. On another note I find it slightly strange, that in almost every threat about abuse-c the topic of data accuracy is brought up, but policy proposals like the abuse-c for legacy space has been withdrawn due lack of consensus. Thanks, Tobias