Tuesday, December 18, 2007
This led to a major effort to rethink the entire supplemental index. We improved the crawl frequency and decoupled it from which index a document was stored in, and once these "supplementalization effects" were gone, the "supplemental result" tag itself—which only served to suggest that otherwise good documents were somehow suspect—was eliminated a few months ago. Now we're coming to the next major milestone in the elimination of the artificial difference between indices: rather than searching some part of our index in more depth for obscure queries, we're now searching the whole index for every query.
From a user perspective, this means that you'll be seeing more relevant documents and a much deeper slice of the web, especially for non-English queries. For webmasters, this means that good-quality pages that were less visible in our index are more likely to come up for queries.
Hidden behind this are some truly amazing technical feats; serving this much larger of an index doesn't happen easily, and it took several fundamental innovations to make it possible. At this point it's safe to say that the Google search engine works like nothing else in the world. If you want to know how it actually works, you'll have to come join Google Engineering; as usual, it's all triple-hush-hush secrets.*
* Originally, I was going to give the stock Google answer, "If I told you, I'd have to kill you." However, I've been informed by management that killing people violates our "Don't be evil" policy, so I'm forced to replace that with sounding mysterious and suggesting that good engineers come and join us. Which I'm dead serious about; if you've got the technical chops and want to work on some of the most complex and advanced large-scale software infrastructure in the world, we want you here.