begin quoting John H. Robinson, IV as of Wed, Jan 19, 2005 at 03:59:57PM -0800: > I have a solution to the munge or not to munge problem: user-definable > reply-to munging. Use one address to subscribe if you want munging, and > another address to subscribe if you do not want munging. This also neatly > sidesteps the problem of setting a default.
Works for me. > The other way is on the confirmation, is to have one confirmation > address for munging, the other for pristine. I think I like this better, > so we only have to advertise one subscription address. We also sidestep > the default issue. Also works for me. > Is anyone interested in adding this to Mailman? Never worked with mailman. No idea what it would take. > ------------- [snip] > > The majority of users expect to reply to the list, so munging is good. > > Everything else below is just so much propoganda. > > You start of with a correct statement (The majority of users expect to > reply to the list) then jump to a conclusion (so munging is good) > without showing how not munging prevents users from replying to a list. r)eply to a post on the list and the email fails to go to the list. Been there, done that, wondered why. > Users can, Users can also telnet to port 25, too. > and capable MUA's do make this easy. Not demonstrated. [snip] > > It's the user's fault, not the list. Don't blame the software for what > > the user fails to do. > > It is about user expectation. The user expects it to go to the single > recipient. It didn't because of the fail-dangerous condition of a munged > list. Only if you expect DWIM software. When I (r)eply to something on a mailing list, I do *not* expect it to go to a single individual. I would rather have a new and different command to _only_ reply to an individual. In fact, that's one of the reasons I like mutt -- it puts the headers Right There where I can edit 'em. > So, do we munge to meet user expectation, or do we not munge to suit > user expectation? Hmm. Conundrum. Until we can come up with mind-reading software, your solutions as above seem the best -- select it when you join the mailing list, or have it as an attribute you can set. [snip] > Not at all. It is about the loss of information, and failure-safe modes > of operation. Don't forget the principle of least suprise. The most catastrophic failure mode is the one where the user is suprised. [snip] > The uncommon case I was referring to is when a person wants a reply to an > address other than a listed From: address. This is what the Reply-To > header is for. And the .signature block, because Reply-To headers have never been reliable. In fact, when it doubt, use the email address in the .signature, and not anything listed in the headers. > I have seen it used for that, specifically in the case > where one person is speaking for a group. The From: indicates the > speaker, and the Reply-To: is set to the group email address. Since that's effectively what's going on with munging on a mailing list, where's the problem again? > I will also note that the two part question, how ``bad things being > easy'' and ``uncommon things impossible'' can be considered better, went > unanswered. Because they presume their conclusion; attempting to answer them implicitly concedes that munging is bad (which is what you're trying to show) and that uncommon things are impossible (which is an unsupported refutation of an assertion). > Personally, the majority of the mailing lists I am subscribed to fail to > munge headers. The vast majority of the mungers fall into two groups: 1) > yahoo groups, and 2) Mailman. A majority of the mailing lists I /have/ been on but have discontinued have been non-munging. So counting up mailing lists seems kind of fruitless -- either we argue the actual merits, or we take a popularity poll. Mixing the two approaches seems fruitless. [snip] > What the majority of people use is not fortunate nor unfortunate to me. > What they use has little to no effect upon me directly. Those lists > that insist upon kowtowing to users of such inadequate software are > making it my problem, however. This puts the blame squarely where it > belongs: on those list admins that believe that munging is good. I actually agree that supporting broken software is not a good thing; better to fix the software. [snip] > I would also be very surprised if Outlook did use mutt's keymappings. You > may submit a feature request with MSFT for them to do that. I use mutt's > keymapping as an example. If someone else's MUA uses different mechanism > to achieve the same end, the muscle memory will accommodate said > differences. There's always telnet and port 25.... -Stewart "Presumably procmail could add in Reply-To headers..." Stremler -- KPLUG-List mailing list [EMAIL PROTECTED] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
