From a programmer's point of view it is pretty simple to code a bot, which scans text looking for certain keywords and then chooses spam text according to keywords found. Nothing special about that. But I wonder how did the spambot pass the captcha - the captcha reports all the false replies in a log. So it seems I have no way of knowing which question did the spambot manage to answer correctly.