Someone who needed an Internet search engine in 1997, would have had quite a lot of trouble finding Google. The reason is that the popular search engine did not exist before 1998.
How did the Google manage to become the number-one search engine in such a short time?
Google's strength is a smart combination of hardware and software. For hardware Google uses no hugely expensive supercomputers, but a network of thousands of cheap PCs.
But what is most important is Google's software - which contains so much intelligence that it is now being licenced to a number of other companies.
If you type in the word sound, you might probably expect to find addresses where you can buy audio equipment, or articles about sound waves and their properties.
What you would probably not expect is an article in which it is related that the harbour of Georgetown should be sounded regularly, neither would you expect to find an article about toothbrushes, in which one of the findings is that regular brushing is supposed to be conducive to sound teeth.
How can a computer programme know that the latter two options are usually not what humans are looking for? A very primitive way to find the relevance of a word in a text could consist in counting the incidence of the search-word in the text. This obsolete technique, which was used by some older search engines, was sometimes exploited by some sites, which had long lists of words they wanted to be associated with, in invisible colour at the bottom of their regular text.
The way in which Google's programmes try to find out the relevance of words in a text, has very little to do with the 'technique' described in the last paragraph.
The algorithm (that's what the way in which computer programmes work is called) of Google's search programmes is called Page Rank. To a certain extent, the principle looks like democratic elections. If many web pages refer to web page X, the page rank of page X will become greater.
But the page rank of page X will also depend on the page rank of the referring pages. For example, page X will rank higher when it is referred to by the web page of The Times than when Tom Luif's Homepage contains a link to it - at least that's what we'll just assume for the time being.
Besides, Google makes use of very advanced techniques to evaluate the relevance of words in a text. Little, however, is made public about the details of these techniques.
Google does, however, not rest on its laurels; continuously Google's engineers are trying to refine all aspects of the search process.
Thanks to the fact that the search process has been fully automated, it is almost impossible to influence the page ranking from outside. Google does not accept payment for higher page-rankings.
