There’s a lot of thinking going on late at night in the labs of various search engine companies. How can they tweak their results ranking so that people get more reliable access to the information they were looking for?
The problem is context. If you search using only one word, eg ‘Chameleon’ are you interested in the creature with the independent eyes, the rock band or one of the twenty or so ‘original thinker’ web design companies by that name?
No search engine can read your mind, so you tend to get dross and then have to repeat your search (rather than scroll through ten pages of results).
Today’s invention is a better way to handle this problem. It would be easier if the searcher could be provided with some kind of tick-box interface, to help specify the context, when first they visit their favourite search engine. To provide a meaningful set of boxes in general is intractable, though. Instead, I suggest the following.
A searcher would undertake a first pass search. He or she would immediately be presented with a spread of results (ideally, one relating to each of the possible different Chameleon entities). The results themselves would simply be the first ten nouns from the front page of each website. These would each then be selectable, by clicking, as a way to refine the next level of search; which would then deliver more of the required data in the usual format. So it’s an automatically generated field of context-sensitive tickboxes.
This would at least limit the madness of being presented with tens of useless results pages.