Currently aerc (or better the go-message) sets the part after the @ in the Message-ID to the hostname of the machine it's running on. This means it's not necessarily a FQDN. In my opinion there are two arguments why this behaviour should be changed:

1. Some spam filters give bad credits for Message-IDs without a FQDN
2. RFC4322 recommends to use a unique domain to ensure the uniqueness of the whole ID. But host names of local machines are usually not as unique as a FQDN.

I understand why go-message generates it this way, simply because it cannot know about the domain of the mail server. But aerc does know the domain name and could easily re-implement the message generation of go-message but with the domain name of the mail server and set it via Header.SetMessageID.

What do you think? Is this the right way to do it? If yes I would also be willing to send in a patch to fix this.