"transactional" report type (was Re: [taugh.com-johnl] Re:
[feedback-report] New version of the ARF draft)
christopher at pobox.com
Thu May 24 15:38:02 PDT 2007
Since there's probably a few folks unaware - spam report handling is
not automatable: 99% of all spam reports (if you're a legit ISP and
not a spammer :-) are just dumb-users (mostly AOL) hitting the spam
button when they *meant* to hit the trash button beside it. Of the
last 1000 or so ARF spam reports I've waded through over 12 months,
100.0% of them were legit messages - asking users why they report them
as spam always yields a "Oh - did I? Sorry" response.
If spam reporting is to be made automatic - users who report spam need
to be given more choices, so robots can handle it properly:
(x) I just want to delete this email
(x) This email came from my ex
(x) I don't like what this email says
(x) This email is an unsolicited bulk advertisement
etc etc - with a second question:
(x) Yes - I want to send this email to several administrators, all of
whom are going to read it and take action to block the sender
It might sound stupid, but loads of AOL folks would no doubt be
mortified if they realized that their personal nudie pics, office
affair emails, business contracts etc are getting broadcast to ISP
admins when they hit "spam" instead of "trash" by mistake - a
*serious* privacy problem.
Friday, May 25, 2007, 8:16:46 AM, you wrote:
JDF> On 2007-05-24 14:21, Damon Sauer wrote:
>> I would like to continue to participate in this forum, but please, if
>> you think I am off-base, or just plain loco, please direct those
>> comments to me (even publicly) but not my employer whom is innocent of
>> the slander here.
JDF> The only part of this conversation where I think you're acting loco is
JDF> in thinking that I'm attacking you or your employer.
JDF> The use case that I thought you were describing is one that's been
JDF> proposed before, so I wasn't exaggerating much -- but I'm glad to hear
JDF> that you're asking for something entirely different.
>> I am not asking that the ~user~ further classify the messages they are
>> marking, but that the format allow for a section that could contain
>> information saying that the ISP's internal systems have checked
>> against their known transactional sending IPs, domains, whitelists,
>> etc. and that the email in question either matched or did not match.
JDF> So, hopefully understanding a little better now, the use case you have
JDF> in mind is:
JDF> 1. [the sender] sends message to a user
JDF> 2. the user reports it as spam to their ISP, via whatever mechanism
JDF> 3. the ISP checks their own systems to determine if the message was
JDF> likely to be transactional
JDF> 4. the ISP reports it to [the sender], using ARF with a report type of
JDF> Seems to me that only the sender can know for sure whether the message
JDF> was supposed to be transactional.
More information about the abuse-feedback-report