Received: from localhost by CS.UTK.EDU with SMTP (cf v2.9s-UTK) id XAA04219; Wed, 16 Aug 1995 23:37:46 -0400 X-Resent-To: drums@CS.UTK.EDU ; Wed, 16 Aug 1995 23:37:45 EDT Errors-to: owner-drums@CS.UTK.EDU Received: from munnari.oz.au by CS.UTK.EDU with SMTP (cf v2.9s-UTK) id XAA04206; Wed, 16 Aug 1995 23:37:42 -0400 Received: from mundamutti.cs.mu.OZ.AU by munnari.oz.au with SMTP (5.83--+1.3.1+0.50) id AA18167; Thu, 17 Aug 1995 13:37:26 +1000 (from kre@munnari.OZ.AU) To: drums@CS.UTK.EDU Reply-To: drums@CS.UTK.EDU Subject: Reply-To again ... Date: Thu, 17 Aug 1995 13:37:07 +1000 Message-Id: <10757.808630627@munnari.OZ.AU> From: Robert Elz Sorry everyone, I didn't really intend to stir up such a debate. I think I agree with almost everything Keith Moore has said on this issue (and Eric Allman, though there's been less of that), perhaps with the exception of the comments on what mailing lists should do, or not do, with Reply-To: headers, upon which I have no strong views (it all depends on how the mailing list sees its role in the message distribution, there is no one true answer). Apart from that I think I see some confusion, perhaps caused by people imagining every possible message context, and use of Reply-To:, From:, etc, and then imagining them all combined into a single message (a program sends mail at the instruction of the secretary of someone, to several individuals, as well as a mailing list that ...) In every realistic case I have examined, the correct settings of From: and Reply-To: can be used to achieve almost any desired behaviour, from the sender's point of view. The recipient of course can choose to send the reply anywhere desired - including, of course, to addresses not mentioned in any headers anywhere. Ideally, as much as it is possible to constrain this, the recipient's mailer should have at least two functions, "reply" which should send to the addresses in, and only in, the Reply-To: header, if there is one, or to some set of addresses from the From: To: and Cc: headers if not (ideally the user can configure the elements of that set that are to be used by default). The second function, Jacob's "reply-personal" should send to the From: address, always the From: address (or addresses), and never anywhere else. A useful third command is "reply-everywhere" which (ignores configuration and) constructs the reply addresses from every address in the headers, on the basis that its easier for the user to delete the addresses he doesn't want to send to from that set, than to add new ones. This would include Reply-To:, From:, To:, and Cc: (but still probably not Sender, though I could be convinced otherwise on that) with duplicates supressed of course. kre