A few months ago, in a blog post, our CTO opined that it was ‘time to return to the Semantic Web again’. As a testament to this, the government has been engaging with Linked Data developers in order to open up data held in public sector departments and publish that data as Linked Data. Whilst still in its initial stages, the discussions on the UK Government Data Developers Mailing List have unearthed some of the obstacles and concerns that need to be overcome in order for this ambitious and much-needed project to succeed. They also surfaced in a recent workshop we ran with a large public sector client, who were in need of some help and guidance in understanding the concept, purpose and advantages that the Semantic Web can bring to government.
One of the concerns centred on the hype surrounding the Semantic Web and whether this really is a step in the natural evolution of the Web or a passing trend that will disappear like so many others have done before. This can easily be answered by the fact that the Semantic Web is not something that has just appeared out of nowhere but , in fact, been progressing for almost a decade.
Another concern was raised about why companies such as Google and Microsoft do not seem to be part of the Semantic Web effort led by the W3C. How do we know that Microsoft won’t invent some proprietary format that trumps RDF and gains widespread use? This concern arises from the confusion between applications labelled as ’Semantic Web’ when, in reality, they are merely ‘semantic’. With the increase in the amount of Linked Data that has been published, it could be argued that a Semantic Web application is anything that consumes or meshes data utilising the URI as the method of entity definition. Even though Microsoft and Google do not appear to be following this mode of application development, they are both actively involved in developing semantic applications. The Bing search engine is using parts of Wolfram Alpha’s API and Google has included support for RDFa with its Rich Snippets enhancements of search results.
The final and most significant concern that surfaced during the discussions was down to not seeing any concrete application that could provide an instant understanding of the problems that could be solved by the Semantic Web. A demonstration of a few SPARQL queries from DBpedia and a glimpse at the excellent visualisations produced from government data were enough to provide some people with an ‘Aha!’ moment were things seemed to fit together and the Linking Open Data agenda clicked into place. A look at some of the quick hacks put together at the LOD-athon also shows the potential of Linked Data based applications and the ease at which they can be put together without having to invest in learning new skills or understanding the inner workings of Web architecture. However, even after a significant amount of evangelising some people were still not convinced that the time and money needed for opening up their data would bring any significant advantage or gain.
Despite the enthusiasm not being shared across the board, there is too much weight behind this project for it not to succeed. Nigel Shadbolt and Tim Berners-Lee are driving this from the top in the UK. At the bottom, you have a small but growing community of developers who are willing to spend significant amounts of time and effort helping those civil servants who see the value of publishing data online. The emphasis at the moment is purely on making the data public (in any format) so that these developers can do the data conversion and create applications that show the value of the Semantic Web to those in the middle tier who are more sceptical. Once those who are higher up have been convinced, the project will really take off and the government departments themselves will be publishing their data as RDF. There is still a way to go but the end objective is achievable and realistic.
It will be interesting to see how quickly the US equivalent progresses in comparison to the UK. It would seem that the sheer quantity and volume of data in the US would make the mass publication of public data a slow process. However, as in the UK, it is often the culture and bureaucracy of a department that are the overriding factors in embracing change, rather than any technical limitations.
It is clear that the government drive will push the Semantic Web to a new audience and, as Sir Tim Bereners-Lee points out, exciting times lie ahead. The challenge now is for the Semantic Web to be taken seriously at both government and enterprise level by people who would have heard the hype but have not understood the detail. Within the enterprise, using Linked Data has the potential for changing the face of Information Management and Data Integration. The newly launched Business Of Linked Data group and our own internal Semantics for the Enterprise Working Group will seek to address this issue and future progress will be posted here over the coming months.