I have an opposite opinion: I don't think it's even logically possible NOT to mess with the results. Let's say there here's a table index and there's some quer

I have an opposite opinion: I don't think it's even logically possible NOT to mess with the results. Let's say there here's a table index and there's some query asking for things the index doesn't contain. How do you calculate "relevancy" out of no-answers? Or let's put it another way: people are asking, the machines are answering. Do you really think a machine is able to answer any type of questions a human person can ask? What would a machine answer to "Who's the most beautiful of them all?" "Snow White" or "the evil Queen"?! Although it doesn't seem, people don't usually want to know existing answers that are already indexed somewhere; they want to know "How will the weather be tomorrow?". Google will answer to that question by looking up words, ranks, links and so on and it'll provide you the most "relevant answer". For Google, not for you. You are left with taking this answer for granted, although it comes out of Google's bellies exclusively. More than that, the answer tends to become "correct" although it cannot bear this value. A search engine that's not already containing the answer (truth-bearer) will only offer opinions (non-truth-bearer); these "opinions" are never true or false, so they cannot be part of any reliable judgement. Google says these answers are to be trusted because they're relevant, not because they're true. But this relevance is not dependent on me, the one asking the question, but on Google's bellies! And you have any doubt that a search engine is designed to be 100% human intervention free?? (here is a more detailed article on this issue: http://utestme.com/post/2745142659/the-search-of-doom http://utestme.com/post/2745142659/the-search-of-doom )