There’s a whole lot of mystique, paranoia, and guessing as to how Google comes up with its generally best-in-class search results. Steven Levy at Wired digs in to discover what really makes Google’s search engine different, and how it learns from us.
PageRank, the generally accepted metric of, among other things, how often a page is linked to, is only a small part of the larger story at Google. Talking to Google’s engineers and tracing the history of publicly announced search features, Levy discovers that a good deal of what Google has learned about search comes from the searchers themselves.
Take, for instance, the way Google’s engine learns which words are synonyms. “We discovered a nifty thing very early on,” Singhal says. “People change words in their queries. So someone would say, ‘pictures of dogs,’ and then they’d say, ‘pictures of puppies.’ So that told us that maybe ‘dogs’ and ‘puppies’ were interchangeable. We also learned that when you boil water, it’s hot water. We were relearning semantics from humans, and that was a great advance.”
If you’re at all intrigued by what Google gets right or wrong, Levy’s piece is well worth the read. It’s a lot of straight talk from inside Google about search, written up in plain English.
