Here is a trivia question for all the minutia lovers out there: what former US surgeon general looks like colonel sanders and has a name that has something to do with chickens?
Go ahead. Type that into Google, bing, or Wolfarm alpha, and see what you get. Chances are it’ll be a long laundry list having to do with Obama picking his surgeon general, KFC, or chicken jokes. In fact, you will get a whole load of matches to your query, but you will not one simple thing: an answer.
This is, of course, because search engines don’t understand questions. They simply scan your search terms for keywords and try to give you relevant pages. They do some rudimentary grammar analysis to try and determine the subject of the query (IE the thing you are actually looking for) but more often than not, they get it wrong (which is why you get more entries about chicken then the surgeon general). Search engines have a hard time with descriptions, too. A ‘man who looks like a thing’ is the sort of thing that a search engine simply can’t handle. And finally, search engines can get easily confused in determining what pages to return that might contain an answer (Consider this post, for example, it has links to search engines, chicken jokes, and observations about grammar. If you had to quantify it by keywords, you’d end up with some mighty odd matches)
I say this because there has been a trend recently of creating “answer engines” – search engines that can understand your question and miraculously supply you with the answer. It started a few years ago with “ask Jeeves” (now part of ask.com) and had its latest arrivals in the much publicized Wolfarm Alpha and the bing “decision engine”. Sadly it seems that hype aside, there really is no noticable difference between a search engine and a answer engine, with the possible exception of Wolfarm – the first engine that has the desency to tell you when it doesn’t understand what you want.
I’m not faulting search engines developers, mind you. Understanding plain-English questions is a hugh and daunting task, and the field is really only init’s infancy. Search engines have gotten a lot better over the past few years, and will continue to improve (and users will continue to get better in searching, which is a different topic for a different day.) But we’re still far far away from the day when all the knowledge of the Internet is at our fingertips. Search engines can fill in many details, but they’re no replacment for a structured approche to learning, no replacement for simple thirst for knowledge, and no good in trivia. At least, not yet.
For anyone who’s wondering, the answer is
C. Everett Koop
a great name, if I ever heard one.