Some UNIXites have complained heavily that modern mailing lists munge the Reply-to: field in the message header to point back to the list software. Mailing lists do this so that replies automatically get sent to the list software. However, sometimes this creates problems, because a Reply-to: header changes the default understanding that one replies to the author of an email. Consequently, occasionally someone will end up sending libelous material to a list thinking she had replied privately.
The argument against reply-to munging essentially suggest that subscribers use the reply-to-group functionality in their mailer, even though most users don't know it exists, and for those that do, replying-to-group is unnatural. The essay suggests unhelpfully that
However, the real problem is that the reply-to-group function is even less correct than normal "reply". Not only will the reply-to-group send a copy of your message to the list, but to the original author as well. This means the original author receives two copies. Things get worse in a long message thread as the To: field grows with each reply, snowballing the mailing list. This is just plain aggravating. To get around it, as a conscientious user, one has to edit the message header in every case instead of only exceptionally in exceptional cases. This is not a HumaneInterface, and so people don't do it.
Really, the essay is just another bastion of the great UNIX user interface design philosophy:
["Reply-To" Munging Considered Useful]
If you are trapped on a list whose administrator is insane, you might want to consider this strategy if you're willing to ruffle a few feathers. Create a Yahoo! group or some other mailing list that has the correct behaviour. Subscribe the Yahoo! group to the insane list (!!). Unsubscribe yourself from the insane list. Read and reply to the Yahoo! group.
In [RFC 2282], the IETF finally clarified what's supposed to be done with the Reply-To field: the list is not supposed to munge it.
The list software is not the author of the message, in any sense. It isn't even the editor. It's the publisher.
["Reply-To" Munging Still Considered Harmful. Really.] has more information on this.
I don't think a generalization and assumption such as this second "law" can be considered a law. I'm on several software-related lists. All use ReplyToMunging. -- EarleMartin
I find people who are religious about this to be really frustrating. All one has to do is look at the actual observable behaviour as a result of doing it the 'proper' way to see how incredibly dysfunctional that approach is, even if it is ideologically correct. It was worth trying, but decades of history have proven it doesn't work, and so one should munge the reply-to. -- SunirShah
It should be noted that many mail clients also have an option to ignore the reply-to header or ask whether to use the reply-to or From before sending the message. -- Anon.
I've moved this to a page of itself, it's an important enough topic. see also http://css-discuss.incutio.com/?page=CssDiscussListHeaders -- I'll probably take stuff from that page & bring it here, since it's off-topic there)