When I asked Adam Bailey,
T>What would you feel about a setup where AOL dispatches the letter anyway but
T>(1) warns the sender that it probably cannot be answered and (2) warns the
T>recipient [by adding text to the top of the body] that it probably cannot be
T>answered?
He responded,
B>Once again, since From and Return-Path might have nothing to do with each
B>other, this won't work.
I believe that Adam's meaning is, "when an AOL customer uses the return
address (RFC822 From: if no Reply-To: is present) of a piece of mail and
sends mail to that return address, AOL cannot know at that point whether a
reply's Return-Path will be a blocked address or not, because the reply's
Return-Path may differ from the destination of the outgoing letter." (That,
in fact, was why I inserted "probably" up above, but Adam felt it didn't
take care of the problem.) Or in other words, from the AOL customer's
perspective, "the To: of my letter letter to you, which could have come from
the From: of your last letter to me, may be different from the Return-Path:
of your next letter to me." On a message passed through a mailing list, From
and Return-Path on the same message are likely not to match each other, but
that isn't the matter under consideration here.
If I have guessed incorrectly, my comments below will require revision but
very likely will still stand. So, as Adam was saying,
B>Once again, since From and Return-Path might have nothing to do with each
B>other, this won't work.
That would not prevent it from working! A problem arises only if, out of
Return-Path and From, one is blocked and the other is not. If both are
blocked, the warning is valid; if neither is blocked, go ahead and get as
epistolary as you like.
This "why don't you answer my mail?" problem affects not only distributions
from mailing lists but also one-on-one correspondence, and for individual
email such as that, From and Return-Path nearly always point to the same
address, let alone the same domain. Even for mail from lists, even if From
and Return-Path differ, it will fail only when one is blocked and the other
is not.
(How many of us run lists where the address to which a user writes to sub-
scribe and the address that appears as the Return-Path in the response,
whether it is hand-sent or automated, are in different domains?)
Third, blocks by people who want to receive replies or subscriptions are
almost never placed against the list or the list's site but usually against
the entire Internet outside AOL, as Adam acknowledged:
B>If you have an AOL account, you can try sending them a message through
B>the AOL mail system, if it's that important.
[Sometimes a very shortsighted user makes up a whitelist, and then even
another AOL account can't write to him or her ... unless whitelisting
applies only to non-AOL addresses and within AOL you can either take
everybody's mail or block specific screen names.]
Now, for when the block is against the entire Internet outside AOL, how much
mail has From XOR Return-Path at an address on AOL while the other is in a
domain outside AOL?
Granted, sometimes an AOL user specifically blocks the list's address as a
substitute for unsubscribing. However, then he or she knows what he or she
did, doesn't want mail from the list, and isn't asking for any, so there
isn't a problem there.
On my list the default for new subscriptions is digest, so From (I assume
Adam means RFC822 From: [or possibly the return address] rather than the mbox
format From_, which almost always points to the same place as Return-Path)
and Return-Path will be in the same domain. Generally I've found that AOL
users sophisticated enough to get non-default settings for their subscrip-
tions also are aware of the meanings of their Mail Controls and they don't
get into these situations.
All told, in cases where the AOL user wants a reply, it must be extremely,
extremely rare for one address to be blocked and the other not blocked, so
the chance of a difference between From and Return-Path just isn't important.
Yes, Return-Path != From is a consideration in most email matters, but
(AOL user is blocking From) XOR (AOL user is blocking Return-Path) is just
too darn unlikely to throw Bob's idea out wholesale.
Finally, the suggested warning to the blocking AOL user can be rephrased
to deal with Adam's caution for the rare case where From XOR Return-Path
is blocked. Since I have never used AOL and don't know their corporate
writing style, I don't want to suggest any specific wording.
Follow-Ups:
References:
|
|
