A newsgroup to which I subscribe gets hit with a particular type of weird spam. The content of each message consists of what at first looks like prose (it is vaguely grammatical) but which on reading turns out to be gibberish. I would guess computer generated. Each message the text is different, I think.
My question is: why? The messages don’t seem to be selling anything, there’s no link to a porn site or whatever. Even if someone were just trying to be a jerk by wasting other people’s bandwidth, why would they bother with computer generated vaguely grammatical text instead of just sending a message filled with random characters?
My guess, either someones bored and created a computer operated trolling machine, or (if java is allowed on your newsgroup) maybe it’s grabbing your e-mail address and needs something interesting to lure you in. (And something that won’t let you know exactly who it is just by seeing the topic).
I’ve seen these pages somewhere. I think I assumed it was some kind of code that cool people knew but I didn’t. The messages are normally a few k in length, so it seems to be too long for anything where the message contents are unimportant. Ergo, I thought it might be some kind of code based on padding out the message with nonsense material (which would be decoded by knowing which letters or words to remove and which to keep). But I’ve no idea really.
Maybe it’s some kind of AI project slowly moving towards sentience based on analysing the content of newsgroup posts. Which would be truly frightening.
I think I know how these posts are being generated, but I can’t remember the name. You start with some body of literature, and (say) two random words. Let’s say your first two words are “The boy”. You then search through your literature for every place the words “the boy” occur together, and see what the next word is in each case. You then pick one of those words, and stick it on the end. So, maybe the most common word after “the boy” is “is”. Your sentence now reads “the boy is”. Now, you repeat the process with “boy is”. Perhaps the word you pick is “it”. You’ve got “the boy is it”. Repeat, this time searching for “is it”. Etcetera. For more coherence, you can do this at a greater depth, searching for three or more words at a time.