On Saturday, May 03, 2003 5:45 PM, Vernon Schryver
[SMTP:vjs(_at_)calcite(_dot_)rhyolite(_dot_)com] wrote:
From: "Eric D. Williams" <eric(_at_)infobro(_dot_)com>
However, that's not the main problem. Users do not like setting up
filters. They simply will not bother to configure their software to
accept "transactional" messages but not "ads" from Roving Software.
They will block all or none of your messages. It doesn't matter
whether you use per-message content labelling or you do as Topica has
tried, labelling by domain name or Habeas mark. Unless essentially
all of your messages are wanted, all of your messages will be blocked
based only on the fact that you are the sender (like the most recent
edition of the weekly drivel that your organization tries to send me).
Yes, but wouldn't a recipient want a 'confirmation' message that was
consented
to always?
Yes, but I don't understand your point.
...regardless of message content labels.
Yes, you got my point. The point is you don't need labels and they don't work.
The more specific point was that some messages are explicitly wanted and
ALWAYS, like 'we received your payment and such.
On second thought, I hope you're not talking about the common spam
that says "This is a one time mailing. Unless you respond you will
not hear from us again." That is spam, even when the spammer keeps
its promise to not send again.
No, I am not talking about that at all.
Spam won't stop being spam and won't stop being filtered by having an
additional label.
I agree, but what is the solution for legitimate 'large scale' CRM efforts?
CRM started out with a couple of meanings. In stray corners it might
still have a little meaning. However, in most of the world it is now
meaningless or a euphemism for various bad things including spam.
I agree that semantics change but I think I am using it in the proper context.
If we are talking (on this thread) about ways to 'control' so-called
legitimate bulk mailers. I call that customer relationship management on a
large scale. So what I am thinking is along the line of Amazon et. al. that
send 'opt-in' eMail. On the one hand I like to get notices of books that might
interest me, let's say, on the other hand sometimes it is a burden so I change
my preferences. I don't get your point on that becoming a euphemism for 'spam'
outright (if that is what you are saying) but it becomes MY spam when I don't
want it anymore but I signed up for it in the first place.
What happens to a number of folks (all of us ;-) after that initial opt-in is
the list gets sold around blah, blah, and we start getting spam outta Sri Lanka
and sh*t.
Is
it perhaps end-user categorization at relationship inception?
Those organizations that pay attention to end-user preferences
will--guess what?--honestly solicit and pay attention to user
preferences including whether the end-user wants advertising.
No labelling is needed for messages from those organizations.
I agree, I am not arguing for labels. I don't think labeling makes much sense.
Not only in the context of sender labeling but overall. Messages are labels
of themselves if you are filtering, why would one need any extra bits to make a
filter work -- "I can make a filter just fine to nab your stuff thank you Mr.
Bulk Mailer."
No matter how other organizations categorized end-users, mail, or
anything else, by definition they don't honestly solicit and pay
attention to user preferences. They hide opt-out buttons in fine
print, "forget" preferences, offer "second chances" for important
newsletters, buy target lists from every harvester that comes along,
claim to believe that their customers claims to have pure confirmed
opt-in lists, and the rest of the dishonestly that leads to spam.
I don't consider those legit bulk mailers and management of that traffic would
be subject to other means (I am not arguing for a specific approach, simply
other means).
That has the
possible dual benefit that a) it imposes a 'burden' on the originator to
relate
customer/consumer information to customer/consumer stipulated categories;
and
b) it allows recipient customer/consumer access to the desired policy
configuration, e.g. I control my spam so from vendor A, but all the others
be
damned - I like that idea).
I don't understand that, except that it seems to be based on something
that is simply wrong. "Spam" is never something you asked for or
otherwise solicited.
See above, I am using the term 'spam' to describe mail -- That I don't want to
see/read/receive right now, BUT that I have asked to see/read/receive
previously. So in that case 'spam' is fungible (For clarity's sake I think the
term 'spam' is close to 99% subjective, but can be bounded by a technical
definition, that def. being very hard to come by, but there somewhere.)
That requires labeling per se but it is within the
context of the relational dB of the originator/customer 'system' and not
tied
to messages or its transporting/delivery/rendering system. Don't inject
what
you can not validate as desired (opt-in?).
Well, ok, if that also means "don't send unsolicited (bulk) email."
Yes, in the sense of the Amazon type mail I referred to before. Don't get me
wrong, I don't think any 'thinking' person would sign the dotted cursor to
request any and all kinds of drivel from any company. But I also think that a
lot of the problem is just that, people are not thinking about what they are
requesting from some of these outfits - and when that is not the case the
outfit doesn't give a f*ck about what they said anyway. Indiscriminate and
Discriminating Unsolicited Bulk (or single) eMails can be BAD.
Outfits that are not spammers can and do send email advertising. Ads
and any other mail that is solicited is not spam. It seems likely that
most outfits that are not spammers have databases saying which contact
addresses want which kinds of email, but that's not required and it's
not a concern of the IETF/IRTF. Whether non-spammers don't send by
never sending anything, telephoning before each mailing, maintaining
a database, or some other mechanism is fine, provided it works.
I think we concur, but I have not seen the definitive definition of what we are
dealing with. If it is "people get my eMail address and send me mail I don't
want" that is a very hard problem to solve in the context of Internet eMail,
that is without the obvious draconian measures one must take to eliminate that
type of traffic (on either end of the communication, O or R). I agree and
don't think it is "I give people my eMail address to send me eMail I don't
want" or, "I get eMail that I want from people with my address ".
Please don't talk about not sending spam as if it is complicated and
requires lots of trade rag e-spert acronyms and enormous costs. It's not,
unless your goal is sending only as much spam as you can get away with.
I do not think it is complicated not to send spam, that is not an argument in
any event. Whatever a trade-rag e-spert acronym is I don't try to think about
things in those terms. I try and think of things in canonical terms and not
'what they have come to mean'. I also don't think it has to cost enormous $$,
Yen or whatever or needs to be outside the boundaries of existing technology
for that matter. From what I can sense we have all of the 'tools' we need to
fix the pipes, what we need is a good plumber (or a damned manual) to show us
where to tighten joints.
My sense of spam is as I stated in a previous eMail (just so you know):
On Saturday, March 29, 2003 10:52 PM, Eric D. Williams
[SMTP:eric(_at_)infobro(_dot_)com]
wrote:
... if we are attempting to define 'spam' then
we should not get into a definition of content description. Jim and Dave you
accurately point out that one man's garbage is another man's treasure but I
don't see how this approaches a definition of what 'spam' is. I think I will
give it try.
SPAM - 1. Messaging in the MTS which violates best current practices for MTA
providers to assure proper canonical representation of it's originator. 2. A
message that does not reflect accurate information for its originator or that
is transmitted with simulated information nominally used to trace origination
[that's a tight squeeze as it ignores incorrect configurations]. 3) A
message
with fraudulent tracking information that is in fact flawed at origination to
obfuscate its origin.
SPAMMER - A user, company or other end entity that engages in introducing
SPAM
into the MTS.
I think these may be a start because they do not address intent or content of
the message sent, quantity sent, transport used or who receives it. to me
the
basic problem is you have problems applying any effective filters or blocks
against it because of the improper information used to construct it.
Regards,
-e
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg