Once upon a time on the Internet, you typed a phrase into a search engine: "banana guard" for example. If the creators of the banana guard thought like you, and also used the phrase "banana guard" to describe their site (maybe in the page's meta tags) then a match was made, and you got the page you wanted. Easy.
Later, as the web became a means to make money, nasty people started doing all sorts of nefarious things to make you visit their site. These nefarious things included using misleading terms to describe their site. You might enter "banana guard" into a search engine and find yourself redirected to something far more unsavoury. You could no longer trust a site's creator to describe it accurately.
What to do? Search engines started to employ people to think up ways of describing web sites. It was no longer entirely up to a site's creator to match their site with your search terms. Now it was up to the employees of the search engines. You would find your "banana guard" so long as the search engine employees thought like you. In the increasingly competitive world of search engines, that mattered. However, as the Internet got bigger and bigger the search engines became victims of their own success. If you look at the footer of the Google home page you'll see what I mean. How long do you think it would take you to read and categorise 8 billion web pages?
Search engines got cleverer. They got automatic. They used algorithms which took into account the number of hyperlinks to and from a site. They assumed that if I created a link to your site then I must have felt your site had something of value to offer. Now, if you typed in "banana guard", you would get the banana guard site that the most people had linked to; most likely the one true source of banana guards!
A triumph of objectivity over subjectivity? Perhaps, but hey, isn't it really easy to create a hyperlink? Yes, even automatically. Sure enough, comment spam was born. You should see the number of comments posted automatically (i.e. by a "robot") to AtomicMaestro that link to sites peddling various "medications". The automatic search engines can't tell the difference between highly relevant links submitted by loyal human readers, and the irrelevant links submitted by spamming robots.
I just want my banana guard!
Now, in the wonderfully evolving world of the Internet, a new approach to information management is emerging. I like it. A little computing power, a lot of human input, and all the advantages of scale. Et voila. Folksonomy.
For nifty examples of folksonomy in practice, check out:
- ESP game (plus supporting academic paper)
- flickr
- del.icio.us. My links are here
