Silly filters

The Dom Post reports:

The pornographic connotations of the word “teen” are stopping emails from reaching the government department responsible for youth issues.

The Social Development Ministry is blocking any emails with the word – or its plural, “teens” – from getting through, because it is often associated with advertising for online pornography.

But Labour deputy leader Annette King has labelled the firewall ridiculous, and called for the problem to be fixed.

The word is on a blacklist of terms blocked by the ministry, meaning that only email addresses on a “white list” can receive messages containing those words.

Annette is right – how stupid. Blacklists or filters based on words are almost always over-reaching. Smart anti-spam programmes look at a whole host of stuff to assign a probability something is spam.

Social Development Ministry chief information officer David Habershon said the word teen was blocked because it was often used in advertising for “adult websites”.

“We base our parameters in terms of the words that are on the blacklist based on best practice.”

He would not say what other words were blocked.

The ministry was “continually refining” its list, but had to seriously consider which words were acceptable, he said. “When we find examples of particular words which are legitimately used within certain business units we amend our systems accordingly. But in doing so, we have to weigh up the impact of removing a particular word against the benefits.”

I despair at the thought of departmental staff sitting down and deciding what words to block. Just use a “smart” anti-spam filter that learns as it goes.