What do you think of the article mentioned in the January 2 Medley about how reply-to munging is a bad thing? Agree, disagree? Don't care?

-- Lyn (millett@io.com), January 02, 2000

Answers

When the title of an essay contains the words "considered harmful" (a reference to the famous "GOTO considered harmful" inflammatory essay about programming style), I think it's generally a tip-off that the author has decided that his way of software is the One True Way and everyone else is WRONG. I have little patience for such people, whether they're talking about software design, web site design, or email reply conventions. Normal human beings can learn how to deal with reply-to's that point at the list. AND they have better things to write polemics about.

-- Mike Gunderloy (MikeG1@mcwtech.com), January 02, 2000.

I used to buy the argument, but I've since come around, for a couple of reasons:

There are forums in which you don't want to build a community, in which the discussion should stay strictly technical, and people should, by default, send their messages back to the original users. Although I enjoy reading such lists occasionally, I rarely ever participate in them. They never develop any personality.

No software does the right thing when replying to unmunged addresses. The best case is usually that the sender of the original message gets two messages by default, and if they reply to the wrong one the discussion gets sent off list anyway. The users of my mailing lists screamed loud enough that eventually I had to submit to their wishes and unmunge the addresses.

Finally, and perhaps the nail in the coffin, is that RFC822 specifies that this is exactly what the Reply-To header is for: A somewhat different use may be of some help to "text message teleconferencing" groups equipped with automatic distribution services: include the address of that service in the "Reply- To" field of all messages submitted to the teleconference; then participants can "reply" to conference submissions to guarantee the correct distribution of any submission of their own.

-- Dan Lyke (danlyke@flutterby.com), January 02, 2000.

The text from the RFC822 is certainly interesting, and I fall into the "it depends" category myself. For example, I run a small mailing list that consists of people I know from college. It's easier to set that one up so that replies go to the list. On the other hand, I like the fact that Webdesign-L does the "right thing" and forces people to reply to all to send messages to the list. More often than not, I reply to all and then delete the sender's address and change the list address to the To: field.

My main problem with Reply-to munging is that information is lost. If I want to reply to only the sender of an email, I have to go digging for their email address in order to send them a message. On a small list where everyone knows everyone else, that's not a big deal. On a large list with thousands of people, that can sometimes be a pain.

Anyway, I don't see it as a black and white issue, although I do like to cite that article when I'm irritated by Reply-to munging, like after I send a message to the list that I meant to send to the sender.

-Rafe

-- Rafe Colburn (rafeco@rc3.org), January 03, 2000.

I'm glad you're not one of those one true way crusaders, Rafe :)

I'm confused about the "information is lost" business, though. On all of the lists I'm on, using a variety of software, it seems to be uniform that they show the list address in the Reply-To header and the sender's address in the From header. So it's pretty easy for me to find the individual return address; it's right there staring at me. I suppose this wouldn't work for people whose From and Reply-To are different, but how common is that?

-- Mike Gunderloy (MikeG1@mcwtech.com), January 03, 2000.

I'm of two minds about this. First, I don't think it's a serious problem; second, I think the writer has a point.

In the "serious problem" category, consider that no matter how you organize the mailing list, the mail program, or the RFC, you're going to still reach a point where /the sender has to decide/ where that e-mail is going. That is always going to be the breakpoint.

That said, I believe that we could minimize the likelihood of mis-sent e-mail by tweaking our software. The example of elm/pine's "g)roup reply" is excellent; I don't know of any other software that uses that wording. Still, I'd like the *default* on a list to be "all", so that's not perfect. Probably the ideal would be software that automagically recognizes mailing-list mail and adjusts the default accordingly, THEN asks for confirmation, e.g. "This will go to the entire mailing list. Are you sure?" Listaholics will hate it, of course, and others will get used to something like for sending messages, so you haven't eliminated the problem.

I'll just note here that Pegasus has built-in mailing-list macros; anything I get from a proper mailing list has a clickable line like: [unsubscribe from this mailing list] right at the top. That's for admin-type functions, though. Pegasus also has other behaviors that aren't so positive -- for example, remembering the last way you set the "reply" address (if you check "reply-to" as the source, the next message will use that as the default, when you may want it set back to "to"). That one got me sending three replies to the weblogs list DIRECTLY to the posters, instead.

So, we're back to that original issue. At some point, the sneder has to decide; and whatever the defaults are, s/he's going to goof up on that decision a non-trivial percent of times.

-- Dan Hartung (dhartung@mcs.net), January 03, 2000.

Rafe, get a real mail client! Pine asks me "Use "Reply-To:" address instead of "From:" address?" on Reply-To munged mailing list messages.

No lost info at all, the original sender is still there in the "From:" where it ought to be.

-- Dan Lyke (danlyke@flutterby.com), January 04, 2000.

Information *is* lost if/when the "From:" and "Reply-to:" fields in the original posting as sent in by the sender are different. It's not common, but some people are sort of stuck with it, so it's sort of Too Bad for those people when it happens. It used to be more common than it is. I recall one correspondant who was stuck on the other side of a broken corporate firewall: all his mail had an address in "From:" that was an internal hostname that couldn't be mailed to from the external net. He solved the problem by putting in a "Reply-To:" that gave an address that worked. This was fine until he ran into a mailing list that replaced his "Reply-To:" with itself; people reading his stuff on that list had no way to send a reply direct to him. He had to put a "to reply direct to me, send to foo@bar.com" in his .sig. Not a big deal, but annoying...

DC http://www.davidchess.com/words/log.html

-- David M. Chess (chess@theogeny.com), January 04, 2000.

If the message you're replying to has a different reply-to address than the From: address, Eudora for Mac helpfully gives a low beep to call your attention to that fact when you Reply. I think the idea is that the user will hopefully take a moment, drop out of autopilot mode and double-check that the To: field is truly what they want it to be.

It's a nice little interface detail. -s ........

-- Steve Bogart (bogart@nowthis.com), January 04, 2000.