The Hindu Business Line : Google runs `a little bit like a university'

IF Google has underwear, who will buy it? This was one of the questions that Sergey Brin posed to students of an Israeli high school in 2003. "Hands shot up," writes David A. Vise in The Google Story, from Macmillan (www.panmacmillan.com) .

To Brin and Larry Page, the Google guys, named recently `Men of the Year' by Financial Times, the project was `one of the less technical' ones.

To check, you may use Google to briefly search for the brief, and see a March 2005 posting on http://blog.searchenginewatch.com confirming thus: "UnderGoos (naturally in beta) offers a full-line of Google-branded undergarments." Price, however, is £0.00 on www.google-store.com. "A project that is closer to our hearts is translation," Brin told the students.

The book takes one `inside the hottest business, media and technology success of our time,' to know how Google, a start-up investment of $1 million, became a market value of $40 billion, all by word of mouth.

"With its colourful, childlike logo set against a background of pure white, Google's magical ability to produce speedy, relevant responses to queries hundreds of millions of times daily has changed the way people find information and stay abreast of news," writes Vise in the introduction.

Ten years ago, search was `not pretty,' is how Rajeev Motwani, "a 30-year old professor who had been Sergey's advisor since his arrival at Stanford in 1993", found.

Vise writes about how Motwani had tested a search engine called Inktomi after it was developed at Berkeley, where he had received his Ph.D.

"He typed in `Inktomi' to see what would happen. Sure enough, said Motwani, `It wasn't there. It couldn't find itself." Not so with Google, which shows 800,000,000 results for itself in 0.17 seconds.

"When Google went looking for someone to ramp up its computer network, Larry and Sergey hired a brain surgeon, Dr Jim Reese," informs the book. "Named Google's operations chief, Reese managed the company's burgeon collection of computer hardware."

Vise narrates the story of how the company "cobbled together a virtual supercomputer from cheap, commodity PCs." Back at the Googleplex, the garden-variety PCs got ripped apart, and all `unnecessary parts that would eat up computing power and resources' disposed of, to build streamlined computers, strung together with "software, wiring, and the special sauce that made Google lightning fast."

More than one lakh inexpensive PCs, stacked in refrigerator-size racks, remain `strictly off-limits to outsiders', notes Vise. "PCs burn out and are not replaced. Instead other PCs take over."

Learn about the 20 per cent rule in Google from Krishna Bharat, a software engineer in the company's research group.

The rule stipulates that engineers spend at least 20 per cent of their time, or one day a week, working on whatever projects interested them.

"The 20 per cent rule was a way of encouraging innovation, and both Brin and Page saw this as essential to establishing and maintaining the right culture and creating a place where bright technologists would want to work and be motivated to come up with breakthrough ideas." 3M had something similar - the 15 per cent rule - many years earlier, notes Vise. A success product that had emerged from such a pursuit was Post-it Notes.

"Rather than having employees moonlighting as inventors at home - with the risk that an idea will either fail from lack of resources or succeed to the point that they quit to pursue it full-time - Google gives them both freedom and resources," observes Vise.

When speaking at the Israeli school, Brin had said, "We run Google a little bit like a university. We have lots of projects, about 100 of them. We like to have small groups of people, three or so people, working on projects. Some of them, for example, are related to molecular biology. Others involve building hardware... The only way you are going to have success is to have lots of failures first."

And Brin toys with the idea of plugging into brain `a little version of Google', as you'd know from the final chapter, titled `Googling your genes.'

Dr Craig Venter, who had decoded the human genome, is of the view that genetic information is going to be the leading edge. "Working with Google, we are trying to generate a gene catalogue to characterise all the genes on the planet and understand their evolutionary development. The massive computing power can be used "to analyse vast quantities of data with billions of parts" says Dr Alan E. Guttmacher, deputy director of the National Human Genome Research Institute, cited in the book.

"We are beginning to have incredible tools to understand the biology of human diseases in ways we never have before, and to come up with novel ways to prevent and treat them."

Fabulous read.

All about ADO.NET

SAHIL Malik, who has been working as a consultant in Microsoft technology for about a decade, and also leading the office of Emerging Technologies at the National Cancer Institute, has written Pro ADO.NET 2.0 from Dreamtech Press (www.wileydreamtech.com) . "ActiveX Data Objects (ADO) was the premier data access technology under the Microsoft umbrella before ADO.NET was introduced as an integral part of the .NET Framework," chronicles Malik.

"What sets ADO.NET apart from previous data access technologies is that it allows you to interact with your database in a completely disconnected data cache to work with data offline."

Disconnected data access is crucial for today's high-demand applications, notes the author.

The book has chapters on connecting to a data source, retrieving data in a connected fashion, DataSets, sorting and searching, updating data, and so on. Of value is the chapter on `best practices' where Malik discusses the right tools. For instance, he reminds that data reader consumes less memory than a DataSet.

"A data reader is an object that allows you to access information on only a single row of data at a given time.

What this means is that, regardless of the size of a result set, traversing this result set with a data reader will only ever have a single record loaded in memory at a given time."

There are many flavours of transactions to choose from, writes Malik, listing out the same `in an increasing order of management overhead and decreasing order of performance'.

The list begins with implicit transactions, which are automatically associated with any single SQL statement and ensure "the sanctity of the data during that statement's execution time period," and ends with "storing a snapshot of previous data, which acts as your `recovery contract' and a flag on the `in doubt' rows."

Helpful notes are strewn all over the book. One such reads, "Retrieving a large volume of data within the context of a single connection will always be faster than retrieving small portions and opening and closing the connection each time, because large-block retrieval causes less network roundtrips and incurs less latency."

Useful for the ADO techie.

Tailpiece

"Instead of expelling them we could have used the `delete' option!"

"And then `empty recycle bin'?"

Books2Byte@TheHindu.co.in