Why am I getting all this spam?

From CIT Wiki
Jump to: navigation, search

I don't need pharmaceuticals, and I don't have money in Nigeria!

The first email was sent in late 1971, and it took until May 3, 1973 for the first spam email to be sent. It was received poorly enough that the 'net was spam-free for a couple years following. It was a gentler age...

Oberlin College uses Google and its filtering system for our email. Google sees enough traffic and has the resources to do a pretty good job identifying spam and keeping it from hitting your inbox. But spammers are in this for the money, and so are motivated to evade even the best of filtering systems. It's a coevolving ecosystem.

One person's spam is another's valued communication, however. We can't count on mere machines to be able to identify all spam perfectly! As a rule, it's better to let some spam pass by, while trying not to misidentify valid email as spam. The most obvious spam can be deleted directly, but many questionable messages are placed in Gmail's Spam folder, where you can look to see if those messages from your Uncle Joe are going if you don't see them in your inbox.

Gmail's All Mail folder will contain all your mail, including the contents of your Spam and Trash folders. Unless you're looking for something that got lost, you don't need to be rummaging around in this folder all that much.

Spam is identified in a number of ways. An email contains several components that can help the system sort the wheat from the chaff:

  • The address of the mail server sending the message
  • The application used to send the message
  • The identification of the account sending the message
  • The From: statement in the message, which can differ from the sending account
  • The text of the message itself

Spammers will use whatever they can to evade our filters and get their message into your inbox. They will switch sending servers, forge sending addresses, fake the sending applications, and even hijack legitimate accounts in order to make their spam appear genuine. This is one big reason CIT recommends you use good, hard-to-guess passwords for your ObieID and other email accounts! Spammers will set up machines to attempt to log into your account, working through whole dictionaries trying to guess your password and steal your account. It's likely you have seen the results of a friend's account having been used for spamming.

This brings us down to the fine art of identifying spam based on the text of the message itself. Mathematically, spam filters use Bayesian analysis to spot phrases and sentences that are common in spam messages. The more such phrases in a message, the more likely it will be caught. You have probably seen spammers try to slip past these filters with creative ways of misspelling common pharmaceuticals. Anyone who's worked with Apple's Siri or Google Translate knows how much better the human brain is at understanding meaning than even the best of computer systems. If the filters don't identify the message as "spammy enough," into your inbox it goes.

We do have tools to help Google improve it's algorithms. Using OCMail, when the spam message is displayed, one of the buttons above the message test, looking like a Stop sign or octagon with a "!" in it, is the Mark as Spam button. Marking messages as spam helps train and improve Google's filter. In Thunderbird or other email clients, you can drag spam messages into the Gmail Spam filter, which will do the same thing.

One last reason you may be seeing a lot of spam: Oberlin College manages many email lists. We do this with a separate list server, outside of OCMail. Google needs to trust our list server so it doesn't see a lot of duplicate messages coming from a list and think our server is a spammer. This means that if a spam message is posted to one of the lists, it's more likely to get delivered to your inbox. Most lists limit posting to list members, but even then a hijacked account can be used to send spam through our list server. Most of the time this isn't a problem, but could explain a particular round of spam plaguing you periodically.

Spam, like paper junk mail, will always be with us, but with luck and a bit of work we can keep it to a manageable level. Remember, you can mark spam messages, and the Delete key is always nearby! If you seem to be suffering an overwhelming amount, please let CIT know.