087 – Improving search – 1996

Forward to PC Moments 088
Back to PC Moments 086 – Back to PC Moments Index
Back to Anecdotes

The Star Trek computer doesn’t seem that interesting.  They ask it random questions, it thinks for a while. 
I think we can do better than that.  Larry Page

Larry Page used a computer at home from the age of six.  His father was among the first to achieve a computer science degree and later a PhD at the University of Michigan.  His mother was a database consultant and both parents taught computer programming.  Page himself graduated in computer engineering but took a series of business courses to broaden his experience.

At the age of six Russian born Sergey Brin emigrated to the USA where he attended a Montessori school (as did Page).  His father was a professor of mathematics, his mother worked at the NASA Goddard Space Flight Center simulating how weather might affect space travel.  By the age of nineteen Brin had graduated with honours in mathematics and computer science.  The two met at Stanford where Brin was assigned to give Page a tour of the college.

1996 Brin worked at the William Gates Computer Science Building and co-founded the MIDAS group, Mining Data at Stanford.  This applied attention to the somewhat disorganised Internet.  Early services had attempted to ‘mine’ the information on the Net by matching text or answering questions; Yahoo! had come up with an alphabetic guide.  But none of these could keep up with the Net’s growth.

Page worked on the Stanford Digital Library Project attempting to define a universal integrated digital library.  He routinely used AltaVista to research the Net and learned that it returned the urls of discovered sites but also showed links from pages to other sites.

He realised there was mileage in these links – and investigating this he decided to download the whole of the World Wide Web onto his computer!  Page called this system PageRank and it was the basis of his PhD dissertation in 1996.

Supported by Brin’s professor, Rajeev Motwani, the three realised they had something valuable.  They created BackRub, a search engine that looked back at a web page’s incoming links.

When it came to a proper name for their search engine they brainstormed until one of the team wrote the word Googleplex on a board.  This was shortened to Google and Google.com was registered in September 1997.  A colleague pointed out that it was misspelled and should have been Googol, 10 to the power 100, but Googol.com had been taken!

Google soon reached the point when the bandwidth it used was too heavy for the college.  Increasing numbers of users and searches necessitated more computers and it gratefully received $10,000 from the Stanford Digital Libraries Project.

The Google service was released in 1997 to Stanford personnel as the college url ‘google.stanford.edu’.  It received an amazing response and supplanted all other search engines.

Stanford’s Office of Technology Licensing promoted the idea of PhD students being entrepreneurial and worked with them to patent their ideas and to find commercial partnerships.  A professor suggested a meeting with Andy Bechtolsheim of Sun Microsystems.  He and other colleagues from Sun had developed a business as ‘angel’ investors.

Most people in the early Net business felt that ‘search’ was a minor insignificant tool and not worthy of much time or effort but in August 1998 Bechtolsheim wrote a cheque to Google Inc for $100,000 seedcorn so it could acquire the hardware it needed. 

Google as a company did not exist at the time; it was incorporated subsequent to the meeting in September 1998.  They had to open a bank account to lodge the cheque!  With the help of family and friends the seedcorn was increased to $1m.

The only thing Google has failed to do, so far, is fail.  John Batelle

Google was underway – and it had managed to fulfil two of the PC folklore requirements by starting business in a garage (of a fellow PhD student) and in Menlo Park!

From an early stage the operation was very successful in getting word-of-mouth promotion of its aims and capabilities.  It maintained an attractive and ‘clean’ home page that assisted in achieving a quick response to searches.

December 1998 PC magazine announced Google as the search engine of choice for ‘its uncanny knack for returning extremely relevant results.’  This was not happenstance.  It was achieved by the ranking of links and by organising pages according to the selected font size, style and relative position of terms used.  It was also greatly assisted by a careful blending of its DiY hardware and software.

It was achieved by the ranking of links and by organising pages according to the selected font size, style and relative position of terms used.  It was also greatly assisted by a careful blending of its DiY hardware and software.

By 1999 the two felt the development of Google was interfering too greatly with work for their PhDs; the service was handling over 500,000 searches a day. They decided to sell it off to someone.

They approached AltaVista, a company that was maintaining a greater than 50% share of the search business.  They hoped AltaVista might appreciate the benefits of PageRank and proposed a $1m price tag for it.  But DEC, the owners of AltaVista, were strong believers in supporting only in-house development.

They offered George Bell of Excite@home the opportunity to buy the business for $1m and later agreed with Vinod Khosla to reduce this to just $750k.  Bell could not see the potential – not his finest hour!

In June 1999 Google achieved a funding deal that ensured the two founders would retain control.  Michael Moritz at Sequoia Capital and L John Doerr at Kleiner Perkins Caulfield & Byers combined to provide funds of $25m.

2004 The Google IPO raised $1.7bn, selling 19.6 million shares at $85/share thus valuing the organisation at $23bn.  By October 2007 the shares had reached the value of $700 each.

By 2011 Google was indexing trillions of web pages on more than a million servers located around the world.  Globally it was dealing with over a billion requests each day and has over 700,000 linked sites.

Ranked by Alexa as the world’s top site and of course had that rank for the US too, where it maintained a two-thirds market share for USA searches.  Its subsidiaries YouTube and Blogger also feature in the list of the top ten sites.

But it was never content with just search – its Google Earth has revolutionised our use of maps, Google Sky has us looking at the night sky differently.  Google Streetview gives us views right up to our front door.

Google is for amusement too – try seeking a Googlewhack.  This is a two-word search term that will return a single Google result.  Inverted commas cannot be used because that tells Google to return only those instances where the two words appear side-by-side.  The words must also be ‘legitimate’.  Examples include ‘Italianate tablesides’, ‘storywriters microculture, ‘insomniacs speedwriter’…

Just as experimentation is affected by the observer, so too are Googlewhacks.  There is a paradox in that discovering a Googlewhack brings about its demise. In order to avoid this some claim a discovery by expressing the search back-to-front.  Gary Stock coined the term and established The Whack Stack at googlewhack.com where users can make their claims for a discovery.

A Googlewhackblatt is a version where a ‘one word search’ delivers a singular result.  An Antegooglewhackblatt returns no results.

ForForward to PC Moments 088
Back to PC Moments 086 – Back to PC Moments Index
Back to Anecdotes

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.