Google released Challenges in Running a Commercial Search Engine, and it is quite an interesting read (if you are having a boring day, anyway). Here are the key points it raises:

• The basics behind search eninges - “The Pipeline”:
  Crawling, Indexing, Ranking, Displaying, Serving
• History of the web
• History of Information Retrieval (IR) & What the basic methods were
• Comparing early IR methods with new IR methods (ie those used by web search engines)
• Link Analysis – Hubs and Authorities, and Pagerank
• Search Engine User interfaces
• Uses of IR – IR is everywhere/Everone wants to do IR/Masses use what we do/etc
• Different types of ways to return results

• Stemming
• Clustering
• Image search
• Citation Analysis
• OCR
• Video search
• Content matching (Adwords and Adsense)

Of course, its not as handy as it would have been with the speach as well… But it covers some basics. Nice to get a view as if you were the seach engine…

Download the file: sigir-keynote.pdf