I had a funny surprise this morning in my inbox. A spam email, one more, got passed the (very disappointing) Outlook Junk Email Filter. I expect such emails to be common, since I'm not using a serious email filter, but this one caught my attention. It used the same technique registration systems or comment posting sections use with captchas to prevent bots from filling forms: Images with noise! Here's the image (URLs blurred):
The rest of the email was in a language I do not understand, which reminded me of Tim's rant about GMail, suggesting them to implement Outlook's language/locale encoding junk filtering. Well, guess what Tim? The spam email was encoded in us-ascii with explicit en-US <span> sections. What we need is a filter that recognizes a set of accepted languages, and blocks the rest, not the opposite, and surely not based on the encoding.