On Fri, 18 Feb 2000, Nick Lamb wrote:
> On Fri, Feb 18, 2000 at 12:50:05PM -0500, Glyph Lefkowitz wrote:
> Useful isn't the opposite of Harmful, and I wonder if he realises
> that... The classic paper is "Goto considered harmful", which doesn't
> deny that it's a useful feature but just tells us why we shouldn't use
> it, and I think RtMCH is written with a similar intention.
The main point behind RtMCH seems to be the idea that munging headers will
deny the ability to reply to someone's real address, if someone sets both
a "from" and a "reply-to" header on their own. This "harm" doesn't seem
particularly significant, for the usefulness that it denies. I don't see
the benefit of doing this, except for the fact that it will confuse some
mailers and most people. Why not set your 'from' header instead?
I don't know anyone who currently uses this bizarre feature of e-mail
headers. I am sure there are, but I've been having this same tired
arguement on mailinglists for years :-) and the only people I've ever run
into who ever did that was because they didn't understand how to properly
configure a UNIX mail program. With the advent of more widely-available
POP3 accounts, such people check their email with native Win32 or MacOS
mail clients (which, I might add, don't have a reply-to-list feature).
If you're someone using gimp-devel who would lose out if the headers were
munged, I'd love to hear about it... this might give me a better
perspective on these sorts of things...
> RtMCH notes that it would be a good idea if mailers got smarter, and
> since it was written they have - Mutt and AFAIK the Emacs mailer are
> quite capable of replying to a list (look at this mail, generated by
> Mutt, for an example) without doing group reply...
"the Emacs mailer" is rather vague -- do you mean, VM, mh, or gnus? (I've
tried VM, and I couldn't find this functionality... but then again,
there's a LOT of functionality I can't find in VM...)
Mutt sounds great; RtMCH seems more a paper on the theoretically ideal
mailinglist than a list to which actual people who use a diverse set of
mailer programs subscribe. The fact is that I have gotten at least 2
messages each time I have sent something to the GIMP list, because most
people don't have a mailer which can support this feature, or if they do,
it takes hours to discover and configure: most people don't have the time.
I certainly don't.
This juxtaposition seems especially ironic seeing as how e-mail is
actually pretty poorly suited towards forum discussions like this, and the
reason we use it is mostly because it's easier to cruft list-stuff onto
email than write a whole new protocol.
> The paper you (Glyph) linked says you should demand "reply to vs from
> address" options in your mailer, which breaks the explicit purpose of
> Reply-To:, instead I suggest that you demand "reply to list" features
> in your mailer, and demand X-Reply-To-List or an equivalent. If
> someone knows of an X-Reply-To-List equivalent which is being actively
> used I'd appreciate a private reply about it, and similarly if you
> know a reason not to have such an X-header.
Yes, reply-to-list features would be nice. The fact is that those
features don't currently exist in most mailers, whereas reply to from
address seems more prevalent. Also, the "explicit purpose of reply-to" is
vague -- and in fact, as the paper I linked to said, defined in at least
one RFC as *explicitly for* list addresses.
Since the only purpose I've EVER seen reply-to headers used for is SPAM
and mailinglists, such a feature seems correct to me; I use this feature
all the time. I subscribe to several lists which do munging; none of them
has had any problems (except for one where someone had his mailer filter
out the reply-to header and so he sends duplicate replies to everyone --
this bothers people, but since he's the author of the software the list is
about they don't complain too much). Aside from the counterexample I
mentioned, I never get duplicate mails from then. This is the only
non-munging list I subscribe to, and I *always* get duplicate replies from
Even if you can deceive people about a product through misleading statements,
sooner or later the product will speak for itself.
- Hajime Karatsu