Re: [Asrg] Please critique my anti-spam system

On 2005-01-07 11:30:48 -0500, Michael Kaplan wrote:

No innocent user will ever get an erroneous bounce if a few of the
large consumer ISP adopt the system;


I must admit I find this statement utterly incomprehensible - or more
precisely, I find the idea that anyone of your obvious intelligence
could believe it utterly incomprehensible.  How will - how *could* - "a
few of the large consumer ISPs" adopting your system, or any other,
prevent a small site handling its own mail in, say, San Jose, from
sending an erroneous bounce to a user at a small site handling its own
mail in, say, Ottawa?


I was unclear.  What I meant was that it would be a simple matter for
and email provider to recognize a bounce sent via my system.


And for the gazillion of similar schemes, too? The examples on your web
site don't look like standard DSNs (you don't even say if the "bounces"
are sent with an empty envelope from or with 
<Autoresponder(_at_)domain(_dot_)com>
as the mail headers suggest), they look like free form text. Currently
everybody implementing your scheme is very likely to generate different
bounce messages - and from experience with handling bogus "a mail you
sent was blocked because of a virus" messages I can tell you that they
are not easy to recognize.

The email provider could then check to see if the recipient had
previously sent out an email to the supposed source of the bounce.  If
the recipient had not sent out this earlier email then the bounce that
is coming in must obviously be an erroneous bounce and it would then
be blocked.


The supposed source of the bounce is <Autoresponder(_at_)domain(_dot_)com>.
Assuming you mean the failed address, how does the MTA get that? Is it
supposed to parse the English sentence "The following message that you
sent to Joe(_at_)domain(_dot_)com was blocked ..." (and its Chinese 
translation)?
Please use at least an already standardized format for DSNs (sie RFC
3461) if you want at least a small chance that MTAs will behave as you
expect them to behave.

Now there is no chance that anyone would receive an erroneous bounce.
In your example above the small site in San Jose would still send out
the erroneous bounce, it's just that the small site in Ottawa would
filter it out before the user ever saw it.


More likely the use would see a flurry of such bounces, complain to his
provider, and the provider would block them MTA of the small site in San
Jose completely (and report it to several RBLs). No more erroneous
bounces, but no more legitimate mails either.

If my system is used on a small scale then I would not expect any email
provider to lift a finger to adjust their system to filter out these erroneous
emails.


They will, if they are annoyed enough. But they won't bother to filter
out only erroneous bounces - they will either filter out all messages
generated by your software or all from the host sending the bounces.
Either way users of your software will find that they can't send mail to
some addresses any more.

        hp

-- 
   _  | Peter J. Holzer    | Je höher der Norden, desto weniger wird
|_|_) | Sysadmin WSR       | überhaupt gesprochen, also auch kein Dialekt.
| |   | hjp(_at_)hjp(_dot_)at         | Hallig Gröde ist fast gänzlich 
dialektfrei.
__/   | http://www.hjp.at/ |   -- Hannes Petersen in desd

pgpKm60rZXJ0m.pgp
Description: PGP signature

_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg