The National Spam Database

The premise is that by keeping records of every email sent or received in the UK, terrorist cells can be identified by their patterns of communication. This approach works very well when phone calls are followed (see previous post on the subject). However phone calls are very different from emails.

Phone calls are rarely spam. Marketing phone calls are a nuisance, but the ratio of marketing calls to legitimate calls is nowhere near the ratio of spam to non-spam emails. Approximately 75% of all email is spam. Marketing phone calls come from a few large call centres. Calls from these centres can easily be weeded out leaving only the legitimate phone calls for analysis.

Spam comes from everywhere and swamps the number of legitimate emails received by an individuals. Equally, anyone can be a source of spam. Spam is often sent from home computers infected with viruses. If any analysis is to be performed on the legitimate emails then the spam emails need to be removed from the dataset. Its not a difficult task, the database applies a spam filter.

However, this means that if a legitimate email is detected as spam it is removed from the set of emails to be analysed. Therefore if I wish to hide my email communications I make sure that they are detected as spam. I could do this by including a spam URL in my email. Or compose an email consisting of a spam image and hide my communication at the end of the email. I could ensure that my computer was infected with a virus and was pumping out spam, mixing in one or two legitimate messages as required.

Email communication is a mess, thanks to spam. Using email without a spam filter is now almost impossible due to the sheer volume of spam. If you want to analyse legitimate email communications you need to remove the spam, but if you want to hide your email communication from analysis, you just need to make it look like spam. If you collect all of the nation's email correspondence in a big database, all you will have is a big collection of junk.