Home Picture Galleries Writing Technical Arcana Books Archive About This Site
 

Google search Hamfisted

 
Himeji Castle

Inside Google

There's a very interesting article on Google's IT Strategy on Information Week, talking about their technology, philosophy and culture. While reading it, I realized how amazingly appropriate "Google" is as a name for the company. Every maths nerd worth his (or her) salt knows that the name came from the term googol, which is a number with the digit 1 followed by 100 zeros. This is a truly enormous number, and basically bereft of any practical purpose. It doesn't feature in any common calculations, and really exists only as a Really Big Number to be flung about casually in dinner conversations about the number of particles in the universe.

Google's focus is on the huge, the enormous and complex, which is quite googol-esque, really. Apart from setting up hundreds of thousands of servers in data centres all around the world, they've developed oodles of custom code for massive data sets and distributed computing problems. Much of their time is spent building APIs and toolkits to support working with gigantic gobs of information, with automation of common tasks, performance tuning and hardware optimization all part of the strategy. One of my favourite quotes is "God is in the details" and Google seem to take this to heart, zooming both out to tackle huge tasks, and in to wring a few extra percent out of their hardware, software, people and resources. I'd really like to work for Google (and they have an office here in Sydney) but I suspect I'd have to toddle off and do a Ph.D. in hyperspatial topography to even get an interview.

Another example of a large data set is Google Earth. I'm extremely late to the party I know, but I've been having a metric shedload of fun with recently. Zooming down from the pretty blue ball spinning in the blackness of space all the way to my apartment in Sydney never gets old. It does remind me just how dodgy Australian internet access is though, and how glacially slow the speed of light can be when you're trying to suck large amounts of detailed geographical data down from the other side of the world. I await the creation of the ansible with bated breath.