There is a sort of rebellion on the web these past few months. Google dominates the search engine market, but a lot of people criticize their algorithm, based on keyword search. Because of this concept, some web pages, stuffed with keywords have a better ranking than some more qualitative pages that focused on content quality instead of search engine optimization.
The main issue that keyword-based search causes is that human language is more complex than a simple juxtaposition of words. Some words have double meanings, and if you don't take into consideration the syntax and the context.
Therefore, some companies started prospecting some new ways to search the web, with new concept of searches. Some of them are based on human collaboration (human search engine). The most famous of these one is mahalo.com. I have already tried mahalo for a little while, but I have not really been convinced by it so far. To be honest, I think this style of search engine requires a lot of time to be efficient, and will never been able to be as reactive as a search engine that will go through the Internet on a regular basis.
However, today I have seen another kind of search engine that could be define as a "Google Killer". This is called Semantic Search Engine. As I have said, it is very important not only to check keywords, but also the context why they have been used in a specific webpage. And these new search engine are eager to focus on this semantic aspect that Google do not take into consideration (maybe so far, as Google will probably adjust its search engine within the next few years).
One of them is start-up Powerset.
"Keyword engines treat pages as word bags, indexing their content without grasping its meaning, he said. Meanwhile, Powerset's engine, applying technology developed in-house as well as licensed from Xerox's PARC subsidiary, creates a semantic representation by parsing each sentence and extracting its meaning. "Meaning is what we index," he said." (source: computerworld.com). So far, Powerset focuses on Wikipedia's pages (it is a beta version you can try out). But the fact is that they might be the future of search engine.
Problem that might have Semantic Search Engine
The main problem is that languages are difficult to analyze. Languages are complex, especially because syntax changes from a language to another. Also, because of double meanings, play on words, and all sort of different problem, it won't be easy to set up a reliable semantic search engine solution. However, with the research advances and the will to see Google's supremacy coming to an end, I believe Powerset and like can see a great future before them.
What do you think? Do you think some of these new actors in the market have a chance to challenge Google's search engine model? Do you believe semantic search have a future?
Blogged with the Flock Browser