Will Yardley (12024-12-05): > While I get your point about one-size solutions and "musts" being a bad > thing, I think in this case, the con of foobar here, at least in many > cases, is pretty significant in almost all cases. > > That is, the pro (less risk of copying a single person an extra time on > the same message, or risk of accidentally replying privately when you > mean to reply to the list) affects many fewer people than the con (risk > of accidentally replying to the entire list, with potentially hundreds > or thousands of members, and risking personal embarassment, when you > were trying to reply to a single person).
Ah, but see, you are getting it wrong / strawmanning it here. You configure the list software to set the reply-to on lists like this one where replying to everybody is the norm, because people ask questions, give answers and discuss those answers, all in public. And you configure the list software to not set the reply-to on lists that are for announces and people asking questions or registering to the author. In the first case, the few times where you want to reply “hi, I am replying to you in private to point to you that this person is a crook / that your system clock is badly set / …”, you know you are doing something unusual, you even think it is necessary to call the recipient's attention on it, and at the same time you fix the recipient lists. I grant you that the case of announces mailing-lists is not properly covered by the “always reply to all and let reply-to take care of it” method. There is no way to write in the mail “if the user tries to reply to all, strongly suggest them not to”. A simple user interface enhancement for that case could just be: if reply-to = from, advise user not to reply to all. I blame the people who tried to design the list-reply-to headers for not having designed them that way. So little difference between the bad standard they wrote and a good standard, just because they ignored the human in the chain. > It can be overridden, but (in the case where the list software replaces > it) only if you remember to do so. Yes, indeed. And my point is: if most of the rest is properly configured, then on the few times where you need to override it, you do remember it, because they are exceptional situations that call to your attention. Regards, -- Nicolas George
