Visited Google’s Kirkland campus yesterday. Google graciously offered their space to hold our first PanIIT talk. Peeyush Ranjan, Engineering Manager at Google gave an interesting talk on the problem of Search Engine Freshness.

Abstract
As the web grows in size, so does the amount of data search engines need to process to provide comprehensive results for user queries. One challenging side effect of this growth is the inherent time delay in processing these pages. This results in a problem with web search that all general purpose search engines struggle with today: freshness of search results. Today, a search for [Hurricane Katrina] on Google.com will provide you with many authoritative sources for information about the natural disaster in 2005. That was not the case for days following the actual incident, when the top result pointed to a wikipedia article about 1981 Atlantic hurricane season, when another hurricane by the same name hit Cuba. While the whole world was worried about the damage in New Orleans, Google.com users were being sent to an article about an incident that was decades old. Only with passage of time did web get re-indexed and Google’s results were relevant again. Come and learn about the reasons why it is so difficult to keep the search index fresh, and what we are doing in Kirkland to help one of the biggest challenges facing the leading search engines.

It was an interesting exchange of ideas and potential solutions. Of course, the biggest problem for a solution is to be applied to majority of the pages so that the exceptions can be minimized.

Also, got a tour of the facilities – free massage rooms, Doctor’s room, reminded me of my Luminant days of the extravagant dot com era.