Searching Semantics
May 9th, 2008 David ChungSemantic search has been the focal point of some intense blogging and opinion of late. What is semantic search? Essentially, it disseminates meaning of words and their context within the text to determine whether it is relevant to the search query or not. This is different from the approach given by Google which uses link authority to determine higher ranks.
As widely reported in the tech media in March, Yahoo has begun adopting some of the key aspects of semantic search into their engine: . Amit Kumar explained the shift in the Yahoo! Blog: http://www.ysearchblog.com/archives/000527.html. Through supporting microformats and a variety of web standard friendly practices, Yahoo! Search intends to to use this to increase the quality of information given to the user. An interesting note which I noticed was that fact that Yahoo! will embrace the Dublin Core initiative which has been discounted by myself as just code padding and superfluous to the crawl process given that the HTML meta tagging are already well established. I could be wrong there but now that Yahoo! Search has been opened for third party development anything is possible to improve its search services. Before we know it, the ability to tag and label content on a page will be widespread and will have to factored in on any optimisation process.
Google has recently been noted to have improved their own semantic search routines with stemming (e.g. run vs runners vs running) becoming more sophisticated. Also, recently Google group members noticed that synonyms (e.g sprinter) are now capable of ranking for sites. These observations bode well for the future of semantic search but will Google ever dump its PageRank formula? Unlikely given that PR has been a lynchpin of Google since day 1. Certainly we’ll expect the handling of semantic search to get much better over time
