Is this spam new or only new to me?

Thu 09 October 2008

Recently I've been hit by a series of spam comments that are unusually relevant to the conversation. This is tricky, but given the following patterns, I believe they are indeed spam.

  1. The posts are spread out over history, not focused on recent entries or comments.

  2. Generally the comments are a single sentence, and no more than two.

  3. Googling the comment contents reveals them all to be excerpts from a much larger comments elsewhere.

  4. Each spammer has a single post of their own, a long piece with a weak command of the English language. Googling sentences or phrases does not work on this text. If I had to guess, it was a written by a Markov chain trained on blog posts, perhaps LJ tags.

  5. At the end of a fairly long and poorly written blog post is a text url. Not a link, like one would expect.

  6. None of the users do repeat business.

  7. There have been about 15 of these in the span of two hours.

Item 5 deserves special consideration, because, dear Watson, it doesn't fit. Why would someone go to the length to write a spam engine to register, blog and comment, train it on advanced computing principles, and screw up whatever circumlocutious goal they had in mind?

Either the spammer made a rank amateur mistake, or was afraid Google or LiveJournal would notice links. Both employ automated systems to prevent spam, and might not notice a URL without the HTML linking. Perhaps they serve up some form of malware that would quickly get flagged if directly linked to?

So either LJ caught and stopped this flood, or the spammer caught and stopped the mistake and will be returning shortly. Has anyone else noticed this pattern before?

Comments !