I first heard of ‘Semantic Search’ last year from reading content that had been shared via Google Plus and was really intrigued by the concept. At the time I remember Martin Shervington organised a HOA (Hangout On Air) with David Amerland on his ‘Semantic Search’ book and ever since I have been really interested in how search works.
For those who are not familiar with the term ‘semantic’, the word derives from Greek and is the study of meaning. Linking this to search, this gives us an insight into the information we are directly looking for by using algorithmic formulas which help join the dots on what we are actually looking for by understanding what words mean and how they can be associated with what we are actually searching for on the web.
In the early pages of the book, David enlightens us on how the web used to be gamed or what he describes as ‘The Pre-Semantic Web’. In ‘ye olde’ days of the web, keywords were already linked to the pages that they represented. In the present semantic world, the concept takes out the ambiguity of search by calculating big sets of data.
In the book he mentions that when he was a child he would have a habit of taking things apart and would look at how things function. He does this with Semantic Search by showing us how the mechanics that Google uses.
Above: The three basic elements of semantic search
There are quite a few diagrams in David’s book, for me this one stands out the most as this shows us the components that help drive semantic search.
Firstly, the URI (Universal Resource Identifier) (and yes this sounds very complicated) but in layman’s terms this can be the name of a person or a thing and this is when a web crawler will be used in the process when we are searching for our desired information.
The second component the Resource Description Framework (RDF) translates the retrieving data from where the URI’s are stored and moving it onto another without mixing things up.
To make this point clearer to us, David uses a great example by using geographical positioning. He mentions that his UK address from the name, location, and post code can all be found in a UK database. However, he then mentions in his analogy that what if his data was then shifted to an American database, straight away there would be a glitch because of the post code/zip differences.
However, because Resource Description Framework (RDF) indexes the data through using a web crawler. David implies that the RDF database allows the American database to allow his UK postal address to be processed as the RDF links meaning to the data that has been indexed.
The third part is adding an ontological meaning to the process, the information that has been gathered is then inferred by the address in a town/city and then broken down and categorized by country. The beauty of semantic search is that it is able to pin point our needs through this process.
I particularly used this part of his book as an example because it just shows the way in which search has evolved and also will continue to evolve over time. David’s book has opened my eyes to the way Google’s algorithm has changed (what is more impressive is that he wrote the book in the pre-Hummingbird stages!). The semantic web is a clear game changer and for a business to succeed in these new and exciting times, both trust and a high level of transparency are needed if our intentions are to be ranked successfully in search.