Brad Templeton Home

Brad Ideas
(My Blog)




Jokes / RHF

Photo Pages

Panoramic Photos

SF Publishing


Articles & Essays



The Book




RHF Home

Ty Templeton Home

Stig's Inferno

Copyright Myths

Emily Postnews


Burning Man

Phone Booth

Alice Pascal

The Rules for Guys

Bill Gates


Why "Bulk Mail from a Stranger" is the best definition for spam

Why "Bulk Mail from a Stranger" is the best definition for spam

I've advanced this definition, but not everybody likes it. It turns out that's actually a good thing. Let me explain why.

(To clarify the terms, "bulk" means a mass mailing, even if the individual messages are different, and a stranger is somebody you've never initiated personal interaction with.)

It doesn't cover everything people think is spam

Some of course feel that's a bug, but in fact it's a feature. There are going to be many definitions of spam, but we need one that we can get a universal buy-in on. That's true if you want laws, or even simply TOS rules for ISPs. If you don't get universal buy-in you just get one state regulating one way, and another a different way, and another ISP a different way. This leaves it unclear to people what sort of spam they can complain about, and just pushes the spammers around to places with the definition they like.

To get universal buy-in, you want the narrowest definition that will get the job done. If the narrow definition gets rid of 99% of the junk in your mailbox, it's good enough. And no matter how much wider you made it, it could only get 1% more efficient.

But more simply, with a narrow definition, you can get everybody to agree that what it covers is spam, is the thing we're trying to stop. Oh, they will all have other things they want to add to the definition to meet their own tastes, but each thing one person wants to add that another doesn't is just a cause for infighting -- and all only to get 1% greater efficiency.

Those who want more comprehensive definitions are not blocked from taking steps of their own against that last 1%.

A narrow definition is the only one courts would ever accept

At least in the USA, the constitution requires that laws which regulate speech (yeah, yeah, I know some people think E-mail is somehow not speech because the recipient and sender share the cost. That's for debate elsewhere) must be as narrow as possible to do the job. That's even mostly true for commercial speech.

A narrow definition protects us from doing ill

Our principles require that we never punish the innocent in our efforts to get at the guilty. That's why in a free society, you deliberately design laws and enforcement systems to have cracks that a few offenders slip through.

That's because a perfect enforcement system, that gets every single spam, is surely also blocking or chilling legitimate e-mails that aren't spam. It's just not possible to be perfect. So that means we always seek to err on the side of letting through a small amount of the bad, to make sure we're never, ever punishing the good. We take that approach for murderers, so it's a safe bet that we should take it for spammers.

We simply don't see much mail we don't like that doesn't fit it.

I get 120 spam a day, and it's extremely rare to get one that doesn't fit this definition. If that's true, then it's demonstrable that the other types of "spam", whatever you deem them to be, are not a problem. They are not flooding our mailboxes, overloading servers or ruining the value of mail.

We don't write laws or put in systems of enforcement to solve problems we don't yet have. We don't even put them in to solve hypothetical problems we might have. We wait until we get the problem, and then we solve it, because we have a chance of understanding it.

The definition fits the cause of the problem

We have a spam problem because of mass mailing. Personally written mails, no matter how annoying, can never mathematically be a problem for us. Even a million full-time people writing personal advertising at a rate of one every 5 minutes (and I really doubt you can write them that fast if they have to be individually composed -- no form letters, no mass mailing) could only send 98 million spam/day, which would be about 1 every 2 days to the average mailbox. But in fact, a million people going at it full time with no automation and no boss (because if a boss ordered them to mail everybody it's a mass mailing) is ridiculous. The ease of bulk is a fundamental component to the spam problem.

The other fundamental part is the fact that strangers can put you on their bulk lists, and this can grow without limit, because there are arbitrary numbers of strangers, while the list of people you personally initiate contact with is inherently limited.

Why not "unsolicited?"

While most spam definitions do contain this word, it is inherent that if you solicit mail, you have initiated contact with the sender, and they are not a stranger to you. The word unsolicited can be added if it makes you more comfortable. To be strict, it might make sense to add it if you want to include those very few people who say, "Hey anybody, go ahead and send me bulk mail no matter who you are."

What about spam from companies you do know?

The one major type of mail that concerns some people is junk mail from non-strangers, namely companies you do business with.

Including this in the definition is problematic, because it's hard to figure a way to do that without blocking some pretty legitimate e-mail, ranging from product recalls to billing information to ads people typically want, like notice of a non-free software upgrade.

Fortunately, this type of mail is inherently limited. You just don't deal with that many people or companies. Real spam can come from anybody in the world. Only .001% of the people and companies in the world are not strangers to you.

In addition, you have additional techniques for dealing with non-strangers. Since you initiated contact, you have some control over the situation. If you were a customer of a company, you have economic power. And while you can't opt-out of every stranger's mailing list in the world, it is actually practical to use the opt-out power you already have on this more limited set of people.

Non-strangers tend to honour requests not to mail you again, and they are of course not anonymous, so they can be held accountable. So while I have gotten annoying mail from companies I know, I have always found I could deal with it through other means.

Speech law says that when you can deal with something through other means, you don't use law to deal with it.

Thus I feel this definition is the best one. My studies of spam don't show very much spam at all that doesn't meet it, and the few spam not covered can almost always be dealt with by other means. To me this points to an ideal definition. Strong enough to fix the problem, but narrow enough that all can agree that it doesn't include non-spam.