Many readers of my blog will have heard of Google Code Search and some will remember that I was previously a part of the project. Well, that gives me special pride in announcing the introduction of Code Search into Google's main web search! For a sneak peak have a look here and see if you can spot the difference.
If you recall, it was little under a year ago when Google started going live with universal search. The problem this project attempted to solve was that almost all users of Google only ever make use of the main web search. The solution they had was to bring all the other specialised search results (from Maps, Images, News, etc.) to the main web search. Two major problems present themself in attempting to achieve this:
- The specialised search engines were used to receiving far lower levels of traffic than the web search, but in munging all the results together they would now all have to deal with the traffic levels obtained by the web search they were to integrate with.
- It's all fine and dandy saying that you're going to integrate the results, but deciding what type of results should be shown when is a challenging problem that needs to be solved per specialised search engine, but in such a way that the main web search guys are happy.
Some other examples: printf, atan2, strlen, ClassLoader
How useful will this be? Well, for one I think it will be useful in attracting users' attention to Code Search. I doesn't occur often enough in the results to be overly useful when trying to search for help on some class or function. And since web search doesn't allow for regular expressions, the most powerful feature of Code Search, it is severely hindered in its integrated form. Still cool though in my opinion. Is it clogging up the results page? I don't think so, since it only takes up one of ten results.