Received: from localhost by CS.UTK.EDU with SMTP (cf v2.9s-UTK) id AAA21997; Wed, 11 Oct 1995 00:32:26 -0400 Received: by CS.UTK.EDU (bulk_mailer v1.3); Wed, 11 Oct 1995 00:31:14 -0400 Received: from dogie.macc.wisc.edu by CS.UTK.EDU with SMTP (cf v2.9s-UTK) id AAA21974; Wed, 11 Oct 1995 00:31:12 -0400 Received: by dogie.macc.wisc.edu; id AA14133; 5.57/42; Tue, 10 Oct 95 23:31:09 -0500 Date: Tue, 10 Oct 95 23:31:09 -0500 From: Eric Norman Reply-To: drums@cs.utk.edu Message-Id: <9510110431.AA14133@dogie.macc.wisc.edu> To: drums@cs.utk.edu Subject: Re: What's the Sender header for? > And as for "AGENT", I'd define this as: > + for any message sent as a result of a specific human action, the > AGENT is that human and the Sender address should be a stable address > that *uniquely identifies* that human. Which human action causes the message to be sent? The act of writing it or the act of "dropping it in a mailbox". This is the distinction RFC822 tries to make. The From: header identifies the author or authors; the Sender: header identifies who caused the message to be sent. In just about all cases, the author does both, hence the Sender: header should be missing in those cases. Now hold that thought until I'm done with the main argument below. > + for any message sent as a result of an automatic process, the AGENT > is the human being responsible for maintaining that process and fixing > it if it breaks, and the Sender address should be a stable address > such that mail sent to the address will reliably reach that human. I'm not sure I agree; why would such an address differ from the SMTP MAIL FROM: or Return-Path: address? Hold that thought too. You certainly imply quite stongly that the Sender: address should represent a human being; I can buy that much. Preview of main argument: the Sender header contains a valid (stable, if you will) address. In order for the information contained therein to be useful to me (the recipient), there has to be a reason why I would want to send a mail message to that address; otherwise it's just commentary clutter. > of what use is the kre@munnari.oz.au information to me? > Two things. First, it tells you which particular person was > playing at being postmaster at the time. Why would I (as a recipient) care who is currently wearing the postmaster hat? Why would I want to send mail to that person instead of just to "postmaster"? > and second, and more importantly, > it tells you who my mailer believes really sent the mail. Why do I care what your mailer believes? Why would I want to sent mail to the address that your mailer believes is "really you"? You have already asked me not to do so by putting a different address in the From: header. > I could just as easily stick "From: Eric Norman " > in the messages I send. > I know this is totally pathetic as a security measure, if I really > wanted to forge a message from you, there's no way I would be > sending it via a mailer that added Sender headers, but when > defeating deliberate forgery is not the aim, the Sender header > does add useful extra information, provided the mailers implement > it sanely. Is the argument here that the Sender: header should or should not convey security information? If the former, I'm going to object because (as you said) it's not possible, and I think there is more useful information to be conveyed by such a header. > be sent. The From: header on the other hand, is really little > more than a comment header with syntax rules. Seems to me like identification of the author is quite a bit more. > I know I just sent my other message, and so no-one has seen it > yet, and all will have before they see this one, but to say it > aain with less other stuff preceding... Isn't it a treat to participate in discussions where the participants are scattered in time zones around the world? > I think I prefer the distinction between "real" and "unreal" (or > whatever) addresses as being that the "real" address is one that > the mailer has determined belongs to the entity that submitted the > message, whereas the "unreal" (From:) address is the way the > submitter prefers to identify herself, and could be anything at > all (it is certainly not verified). No more than that. Seems backwards to me (must be that counter-clockwise Coreolis force :). I would say that whatever she wants to be known by is the one that's real. Well, it's clear that I think the behavior of MH regarding the Sender: header is not very useful and nothing more than commentary clutter. So why would I ever want to send mail to the Sender: address (assuming it means whomever caused the message to be sent and understanding that that may not be the author)? I may want to send a message to such an entity that conveys something like: "I consider this junk; please don't sent me any more", or "I have moved; here's my new address", or "I don't do this any more; I'll forward this, but future such requests should go to flyswatter@ayres.rock". I'm tempted to say that I want to talk to the sender when talking about envelope information and I want to talk to the author when talking about the content in the body. Ergo, if a list expander is going to put in a Sender: header, it should represent the person that adds or deletes names from the list. This may or may not be the person that fixes problems. Assuming that the Sender: address differs from the SMTP MAIL FROM: address, I think the latter is mor appropriate to use to try to get problems fixed. -- Eric Norman