I wrote:
Could someone reiterate the thinking behind subscribing yourself to
mailing lists under an alias?
Sean Straw replied:
Well, one thing is that for lists which are archived to web, which in turn
are trolled by spambots to collect addresses, a list alias is easier to
ditch and use something different when the old one is overridden with spam
(at least, this is the case for people who have TRUE aliases, at the MTA
level, rather than MDA - such as plussed addresses).
I can see that being handy, as long as addresses are in relatively good
supply. I just discovered today that my ISP's mail system supports
plussed addresses, so it looks like this is another possibility for me.
("psmith+`date`", here I come.. just kidding).
There's a nifty concept... "one time addresses" (like one time passwords).
Post with it and then throw it away... Not terribly practical, though.
My big issue right now (if you haven't been reading lately) is determining
if an arriving message came from a mailing list or not, because I intend
to apply spam filters differently to list mail vs direct mail.
Each mailing list generally has it's own unique characteristics -
"Resent-Sender:" or whatever. This is the long-standing way of filtering
list mail -- identify the characterstics the list mail you're subscribed to
falls under, and filter it that way.
I'll probably end up doing that anyway. But without some method which
covers all lists, I'm constantly at risk of subscribing to a new list and
forgetting to add a recipe to catch it. Combine that with autoreplies,
and the result isn't pretty.
Perhaps run the basic filters, then filter out lists, then run the
exhaustive filters on what is left.
That's basically the plan.
"psmith-exploder" alias. Also in the scheme somewhere is the requirement
that I have to somehow keep the aliased address out of the distributed
copies of my postings which go to the other subscribers.
I don't follow - you're going to subscribe with the alias, but continue to
try to post using the non-alias? Some lists won't LET you post if the
address isn't on the subscription list.
Not exactly. Era has cleared this up for me recently. You basically
subcribe both addresses to the list (the original and the aliased), but
you ask the maintainer to set the subscription of your original address to
"post-only". Then you just make sure to only post from your original
address.
That way you only get one copy of list traffic, and hopefully your
aliased address appears somewhere in the header allowing you to segregate
it from all direct, non-list traffic (which would include off-line replies
from list readers, non-list spam, etc. You could then "safely" apply
strong filtering (password protection, etc) to mail destined for your
original address, without worrying about autoreplying to someone who wasn't
sending to you in the first place.
So far my biggest problem with that has been that I don't know if the
aliased address will even appear in list mail I receive.
If your mail server adds "for psmith" in the received line, then you have a
way. You're boffed though when multiple people at the same domain get list
mail - the MTA doesn't add it then.
So far it I appear to be the only one, since I do see the "for psmith..."
in the recevied lines. But it worries me that one day that might suddenly
disappear from the headers because someone else at my ISP signs up. In
that case, my filter would start autoresponding to list traffic before I
noticed and shut it off.
scanned through some mail from this list that I have saved, and looking at
full headers, can't spot my own address anywhere outside of the Received:
lines, which I consider unreliable.
Try scanning the post you just made. With a non-exploder address, which
would be the reply-to if I were replying offlist or carboning you. So what
again whas the purpose of the exploder address if people still see your
regular address, and carbons would go there?
Mail sent to the list would look like it is TO the exploder address -- I
would disable autoreplys to those messages. Carbons would get autoreplies
if it's the first time that author sent to me directly. I guess there's
another reason to not carbon someone who's on the list if you're sending
to the list anyway.
This can be avoided to some extent by setting up the whitelist so that
anyone who posts to the mailing list gets put on the whitelist (list spam
and all). That way, as long as the author who is carboning me has sent to
the list before, they won't get an autoreply from me. I might even introduce
a hold time (perhaps 30 minutes) while I wait to see if the cc copy I got
also appears on the list, so that even if it is the first time that poster
has sent to the list (at least since I've been subscribed) he will still
get whitelisted.
Man, this gets complicated pretty quick...
And then there is the issue of "plussed" addresses ("psmith+exploder"),
but I don't know if I'm mixing two different strategies mentioning that
here...
If your mail provider supports plussed addresses, they're useful. However,
unlike an MTA alias which can be deleted, and thus cause mail to bounce,
plussed mail continues to deliver to your account, where you have to filter
it to kill it.
True, if you're using the plussed addresses as public addresses that you
will later throw away. In my case, the plussed address should be only known
to the list maintainer (assuming the subscription list is private), so
hopefully I won't need to change it that often, if ever.
Please DO NOT carbon me on list replies. I'll get my copy from the list.
I think I can appreciate why...
Paul