-
Google Calls Yahoo-Microsoft’s Explanation Of Search Scale “Bogus”
One thing Microsoft and Yahoo were touting as a major advantage to them teaming up, is that their search scale will grow tremendously. The benefit of the search scale, or search volume query, growing is that Microsoft and Yahoo will have more data to analyze and use in order to compete against Google.
News.com interviewed Google’s chief economist, Hal Varian on this topic. Varian called that logic “bogus,” saying that search scale does little, on a statistical base, to help improve search quality. Varian explained why in three points:
- Statistically, Varian said, that there is a “small statistical point that the accuracy with which you can measure things as they go up is the square root of the sample size.” Meaning, statistically, more data at this level, doesn’t really change how you measure things.
- Queries at all search engines are growing approximately 40 percent a year. So there is a huge increase in search scale, even without this deal.
- When Google tests quality and other factors, they do so on 1 percent or 0.5 percent, that is all they need to measure on. So Microsoft or Yahoo can increase that sample percentage size to 2 percent, if they like, and get similar data back.
These are extremely interesting points that I personally did not think about, but indeed make a lot of sense. If you think about it, Microsoft has about 5% share (approximating) and Yahoo has about 20% search share. Microsoft grew tremendously with this, but if Varian is right, then that growth is not as significant, in terms of quality, as Microsoft wants us all to believe.
See more articles by Barry Schwartz >
Google’s Working On New File System
We often don’t think about the core file system and servers that power Google’s properties, from Search and Gmail to YouTube and Wave, but Google does. The Register reports that after ten years with the original Google File System, GFS for short, Google is working on a version two to replace it. The Google File System is different from the underlining Caffeine infrastructure update, as I understand it, Caffeine runs on top of the Google File System, like you run Microsoft Word on Windows.
The original GFS wasn’t built to handle applications such as Gmail or YouTube, but that is exactly what it has been powering for the past couple years. It was originally designed to be the platform to power search queries and is now being asked to do a lot more. Version two, will take this into account. The release date is unknown, but The Register goes through all the technical details that would make any system administrator raise their eye brows. Personally, I don’t fully grasp these details, but conceptually, it make sense.
0 comments: