Last week, I attended the Talis Platform Open Day which brought together practitioners interested in Linked Data technologies from many different industry sectors. Even though the use and creation of Linked Data is still in its relative infancy, it is clear that many people who understand the concept and have published Linked Data themselves see this as the natural evolution of the Web. The ability to name entities and objects with URIs and utilise the HTTP protocol for retrieving data appeals to those Information Architects and Web developers who feel constrained by traditional databases and data storage systems that hide information in silos within the ‘deep Web’. The Linked Data in Action demo showed the value and potential of opening up relatively small sets of data to be linked on the Web. However, apart from a small group of ‘evangelists’ who take every opportunity to hale the virtues of Linked Data, the more business minded and traditional enterprise architects have been hesitant, and even sceptical, of what they see or understand as The Semantic Web packaged up with a different branding.
There are those who still believe that any technology that comes under The Semantic Web umbrella involves complex ontologies, definition of every concept that could possibly be used by the business and a top down approach to data integration. This may have been the case in some specialised domains in the past, but the case with Linked Data I would argue, is different. A mantra for The Semantic Web, proposed a number of years ago went something like ‘a little semantics goes a long way’ and this is especially true with Linked Data.
An incremental approach to Linked Data adoption, initially using lightweight vocabularies to describe the important entities in a business and publishing this according to Linked Data principles, can bring immediate benefits in realising exactly what knowledge the business holds. This is why the linking government data initiative has focused on releasing and opening up data and documents for conversion to RDF in its initial phase. Fitting that data in with other datasets in the cloud and creating ontology driven applications are processes that naturally follow on from the initial ‘opening up’ stage. There is real business value in publishing Linked Data as the process will involve an analysis of what data the business has, who owns it, who controls it and more importantly, what it means.
Another confusion that arises when talking to people about Linked Data is the presumption that they will have to publish all of their data whether private or public. The fact that you are simply making the data that already exists on a web page machine readable comes as a great surprise. What comes as an even greater surprise is that publishing RDF or RDFa data for your web pages can greatly improve your search engine ranking as evidenced by Best Buy. The biggest shock then comes when you explain how easy it is to adapt existing web sites to incorporate semantic data and publish as Linked Data. Since the Web is often the driving force in new technology take-up, this factor alone should make every business start to wonder at what they are missing out on.
Once the initial data publication method has been established the benefits then become two-fold. Firstly, you can begin to understand your own data and how it fits together and work on integrating that data with other systems in the business. Secondly, the publishing of that data increases your visibility on the Web and allows others to mix your data to produce applications you could not have created yourself.
As pointed out at the Talis Open Day, the sites publishing data and becoming part of the Linking Open Data cloud has mirrored the early development of sites that started publishing pages to become part of the Web. In the future, just as every company, big or small, naturally feels the need to have a website, they will also feel the need to publish the data behind the page and join the Web of Data.