ietf-asrg
[Top] [All Lists]

Re: [Asrg] Nucleus of a draft BCP on filtering

2003-03-07 20:43:15
Keith Moore wrote:
Guiding General Principles: - spam is UBE (unsolicited bulk email)

I don't think we're going to end up with a single uniform definition of spam.

I believe we have to, insofar as we need to limit the scope of what the "S" in ASRG means. Of course, some people will call the nagging email from the SO, about forgetting their lunch, spam. But that's just plain wrong.

If we don't limit our scope, we'll just be in an endless rathole.

The thing we're primarily concerned about is that activity/behaviour which endangers the future of email. The behaviour that turned striker into a smoking crater.

Sending 1:1 email isn't it. Having a salesman for widgets pick your email address out of a mailing list about widgets and suggesting "you might be able to solve the problem you asked about by buying our XYZ widget" isn't it. Changing your mind about email you explicitly asked for isn't it. Your neighbor's emailed porn from Germany isn't it.

Anti-spam lobby organizations, such as the various branches of CAUCE, have, I believe, now all fixated on "UBE" as the spam they're concerned about (regardless of the "C" in the name) - I'm a director of CAUCE Canada. "Usenet spam" isn't content. It's simply "the same thing posted many times" [I wrote the consensus standard...]

Perhaps it would be nice to exert effort on a classification system that provides more flexibility than a simple UBE definition. But you can forget about individuals classifying things properly. And we can forget about getting very far. And if we include that, the pro-spam side, and the media will just continue on their merry way "they can't even define it".

Think about most the strategies that people are talking about here. Counting bounces correlated to senders. Establishing authorized "sending MTAs" for given "from" domains. DCC. Spamtrap analysis. Bayesian, Razor, Brightmail recognizing messages that "look like stuff seen before" Etc. That's _all_ about bulk.

The original direction of defining spam as "UCE" (actually, UBCE) derived directly out of a belief that C was easier to legislate given US jurisprudence. Thinking has shifted since then, to that believing a content-neutral definition of spam is (a) easier to legislate and (b) easier to implement (because you don't have to write AI engines to recognize whether something is commercial).

You can automate bulk detection a lot easier than commercial detection.

Yes, some people will call commercial email spam. But why? Why should they care if their neighbor is getting commercial email that they asked for?

Defining the "S" in ASRG as "UBE" helps us focus on solutions that may prevent more strikers. Isn't that what all of us are primarily interested in?

Besides, I was hoping that the _rest_ of that original message would generate the majority of comments ;-)

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