I read an article around schema-less database which sounds cool. (http://bret.appspot.com/entry/how-friendfeed-uses-mysql)
But what isn't clear to me is how do they run search queries on this data? Since the data is in JSON format how do we look for it?
They're using sphinx for that