On 6/4/20 9:10 AM, Laura Atkins via mailop wrote: > What is your mechanism for that trust? If the answer is “someone will > figure it out” then there’s no point in even suggesting such a header.
Well, I would disagree with that, actually. Much of this is automated on the recipient end at large receivers, and if it so happens that messages from a certain large ESP that contain a certain header are only half as likely to be manually flagged as spam as other messages from that ESP, that would be likely to show up in a bayesian/AI classification system. In general, it's in the interest of senders to provide *any* clues they can that certain messages they send are more likely to be legitimate. It doesn't need to be a header; a large ESP signing COI messages with a different DKIM selector would be a similar clue that could be used. Differentiation of mail stream types helps everyone. Even on the human side, it seems people spend vast amounts of time looking for clues about mail they can use to filter it -- I've seen people trying to guess whether certain ESP mail is COI or not based on the sending IP address ranges. If people are willing to do things like that, I think they'd be willing to use any other clues reputable senders provide. But I may be wrong. >I will also say there is such a thing as spam to a COI address. Yes, definitely. Doing such a thing would be unhelpful to the sender, because it would make systems trust the "clue" from that sender less than they otherwise would. -- Robert L Mathews, Tiger Technologies, http://www.tigertech.net/ _______________________________________________ mailop mailing list mailop@mailop.org https://chilli.nosignal.org/cgi-bin/mailman/listinfo/mailop
