BBC Radio 5

Matts on 2002-08-16T12:42:19

So tomorrow I have to do a live interview on BBC Radio 5 Live. This is kind of interesting, because according to my boss, radio 5 are also going to have on the other line a spammer (wonder if it'll be Ronald Scelson).

So I need to arm myself. I know the spammer is going to throw out the usual crap about how its the same as junk mail through your letter box (which it's not), and stuff, but can anyone here think of anything that might be worth bringing up? The one thing I don't know enough about is how spamming in the UK relates to the data protection act. And I bet I get asked about that!

If anyone has any good ideas on what sort of things to talk about, attach them to this journal.


When?

davorg on 2002-08-16T13:46:32

Can you tell us when it will be? I'd love to listen in.

Re:When?

Matts on 2002-08-16T14:00:49

Oh yeah, sure. 8:30pm on Saturday 17th August. Apparently.

I'll confirm with more details when I've actually spoken to the lady at the bbc. You can get a Real feed from here if you're not in the UK.

Online?

Dom2 on 2002-08-16T13:55:42

Any chance of an online streaming version? or have the beeb stopped doing oggs now?

-Dom

Topics

ziggy on 2002-08-16T14:39:18

I'm not sure what standard topics you're going to be mentioning.

When highlighting the differences between snail mail and email, it might be worth talking about how email is more like a direct link into your brain, while snail mail is not. (I first heard that from gnat, from his "be an advocate, not an asshole" talk at YAPC).

Might want to read this article...

Purdy on 2002-08-16T15:54:08

Just out in NewsWeek

Re:Might want to read this article...

Matts on 2002-08-16T16:05:07

That's the guy.

Update

Matts on 2002-08-16T15:59:09

The "spammer" is one "Alan Ralsky" - from the US, claiming that he fully supports opt-out (hah!) and is a legitimate business mailer.

The segment should be about 8:45pm Saturday.

Paul Graham on spam

ziggy on 2002-08-16T18:03:17

The ever-sensible Paul Graham has a new article on accurately filtering out spam using Bayesian probabilities. He's claiming missing 5 out of 1000 spams, with zero false positives. The problem with his technique is that it needs to be trained to see what kind of messages and spam you receive. The benefit is that the probability model is finely tuned to the messages you receive.

There's lots of analysis about spam in the article, including a few well-reasoned explanations on why spam exists and why spam is bad:

All along the spectrum, if you restrict the sales pitches spammers can make, you will inevitably tend to put them out of business. That word business is an important one to remember. The spammers are businessmen. They send spam because it works. It works because although the response rate is abominably low (maybe 15 per million, vs 3000 per million for a catalog mailing), the cost, to them, is practically nothing. The cost is enormous for the recipients, about 5 man-weeks for each million recipients who spend a second to delete the spam, but the spammer doesn't have to pay that. Even so, sending spam does cost the spammer something, so the lower we can get the response rate, the fewer businesses will find it worth their while to send spam.

Re:Paul Graham on spam

Matts on 2002-08-17T08:17:32

Bayes is very good if you can tune it to your type of email. By the looks of things, Paul Graham gets very little business-like emails. That's where we found our largest set of false positives with it. He's also right - doing bayes against word pairs is better than against single words, but your database does grow a lot larger.

We're getting about 90% accuracy with it - on real customer emails.

talk to the interviewer ...

ask on 2002-08-16T21:50:33


If at all possibly, be sure to ask what the interviewers story is before it goes on. He or she will likely then tell you a bit about their story and you can make sure to prepare your thoughts related to that.

How did it go?

chrimble on 2002-08-18T16:39:09

Repeating "I must remember to listen to this" again and again didn't have any effect and I completely forgot to tune in... How did it go?

8)

Re:How did it go?

Matts on 2002-08-19T07:03:08

I think it went fine, but then I could have sounded like a complete twat and my wife wouldn't have told me ;-)

It was a very short segment, especially due to all the news about Holly and Jessica that they were trying to squeeze us in between. But I'm sure my company will be pleased, since the closing words were from me and basically a big advert for our service ;-)

My view

Odud on 2002-08-18T20:04:46

I suspect the amount of time got cut down dramatically - because of the tragic events elsewhere. I wasn't sure what point the Beeb were trying to get across... I think you performed very well - I guess you were a bit nervous and the slightly crackly phone line didn't help (oddly the spammer - from the US I assume - sounded a lot clearer). I think your estimate of the number of man years wasted deleting spam was a good one - of course the spammer made the usual points "we don't spam anyone who hasn't asked for it" and "there's always the removeme link" He was typically evasive about where he gets the addresses from... Interesting that he quoted the American constitution as a defense - I assume he only sends spam to USA users then .

An afterthought

bart on 2002-08-24T08:51:10

I only thought of this a few days after the show (which I didn't hear, BTW).

The main reason we get spam, is because sending e-mail is free. If everybody charged a little bit of money of money for each e-mail received, say 1 cent (be it dollar, Euro or Pound) per recipient, then spamming would cease, simply because the costs would be far more greater than the generated income.

This shouldn't be real money, in most cases. ISP's could allow a certain maximum of mails per month per user. Exchange of mails between two hosts, when reasonably balanced, would effectively be free of charge. Only for those people sending millions of mails a day, this would impose a real cost, up to millions a week. Ergo: they'd stop, almost immediately. If they don't pay, their mail simply gets blocked.

I can still see this work even if the price is several orders of magnitude lower.

The real unsolvable problem I see is for mailing lists. Those would be in trouble.

Just a thought, really.