I remember reading an article back in 99 with the headline "How the Web Shrinks as it Grows". The basic idea was that despite more and more pages getting added to the internet, regular users are visiting fewer sites.
At the time I thought this was due to the high quality of those few sites, with good brands and updated content, etc. But now I think the search engines have a lot to do with it.
Despite there being eight billion plus pages on the web, I sometimes feel that I am watching a cheap cartoon, where the background (the same sites) get repeated again and again. I know there is much more information out there, so why do I keep finding myself on sites and thinking "oh, this one again"?
Take weblogs as one example. There must be over a million of them but I seem to end up on the same handful of blogs all the time.
This may reflect my search habits, or my preference for sites of a certain style and quality, but I think the nature of how search engines reward -- via their infamous algorithms -- is at least partly to blame.
Maybe Google should offer a "randomise the results" button to replace the entirely useless "I'm feeling lucky" button?
Or Maybe Google's regular results should be made more like its News results, where results are clumped into stories, and users can choose from different sources of the same story or piece of information? (This would admittedly be tricky, as not all types of information are as interchangeable as newspaper stories are, but may work with some imaginative thinking.)
Just my €0.02 for today...
Comments
2 comments
Search engines provide an element of navigation to users. Consequently the best optimised will be the most visited. But the element of webrot tends to reduce the page count, or more precisely the positioning of webpages in search results. When you group a country's webpages based on the age of the index page, a minority of pages show up as having been updated in the current year. This is probably down to the bulk of the web being static brochureware site with very little actively updated content. If you are looking for actively updated content then you will narrow down the set of potential websites even further.
Clumping regular results could be a move towards this grail of a Semantic Web. But in reality it would be integrating a directory like categorisation at a fundamental level of the search engine. SE's like Google that are based on linkage type algorithms are at a major disadvantage when it comes to this kind of highly structured directory style of search results. And when you throw blogs into the mix, the idea of including blogs in a standard search engine index is terrifying when it comes to applying any kind of categorisation algorithm. Humans tend to categorise sites better.
A simple "order by last modified date" button on Google's ordinary interface would be very useful.
Interesting idea John. Maybe Google could offer a few different "sort by" toggle options?