Postman Pat Vs. The Internet
Over at Gweilo Diaries, Conrad
Posted by: Pixy Misa at
05:24 PM
Processing 0.0, elapsed 0.0034 seconds.Something needs to be done and, in the end, the only solution may be an e-mail "postage fee".
This suggestion has been floated before, and has been largely ignored because, for a variety of reasons, it is completely impractical.
First off, the Internet is global. Unless every country in the world charges an email postage fee, any country that doesn't charge such a fee will become an instant spam-haven. So the spammers will relocate their servers at minimal cost, and spam will continue unabated.
Second, no-one runs the email system. Anyone can run an email server; I run three myself. Indeed, I've written an email server myself. How are you going to enforce this postage fee, when the way email actually works is one (privately owned) server passing the message to another, with no "post office" of any sort involved?
Third, even if you passed legislation that all SMTP (the Internet mail protocol) transactions on the public internet incur a fee, and enabled law-inforcement agencies to go after the free-email offenders, the immediate result would be that people stop using SMTP and start using something else. It's quite easy to send email over an SSL-encrypted HTTP connection so that it looks just like a web page. Tax that.
Fourth, there are many, many useful public mailing lists that send out thousands, sometimes tens of thousands of messages a day. An email tax would kill them instantly to no good end.
Finally, the technological solutions do work. I have 600 spam emails in my Junk folder, trapped there by Mozilla's Bayesian filtering. Christopher Caldwell's article shows a basic lack of understanding of how Bayesian filters work:The primary tool that exists today is the "Bayesian" filter, which seeks out words like "Viagra" and phrases like "online gambling." Spammers have long been able to evade such filters with subtle misspellings (TURN HER ON WITH HERBAL VIARGA!).
In fact, this is precisely the problem that existed before Bayesian filtering, and which Bayesian filtering is designed to solve.
The key here is that spam looks like spam. With Mozilla, there's a training period where you need to tell the program this is spam and this is not spam. It quickly learns to recognise the characteristics of spam; not just individual words, but all the patterns found in both the headers and the body of the message, the same things that let you tell at a glance that a message is spam.
Which is not to say that I don't favour anti-spam legislation. Even when it's filtered out automatically, I'm still paying to download the spam in the first place. The right legislation would let spammer's internet connections be blocked promptly, preventing the flood of messages going out in the first place... And leading us back to my first point. But at least we won't have some ghostly beaureacracy monitoring our emails and extracting a penny a piece.
Comments
16 queries taking 0.0028 seconds, 7 records returned.
Page size 5 kb.
Powered by Minx 0.8 beta.