At the 09 Online Information Conference there was a strand called “The Semantic Web, Coming of Age”. Mike Powell, Dejan Dincic and I attended for IKMemergent. We came away with a clear sense that there are some significant developments taking place and it is an issue that will impact the people and organisations we work with in Development. I wrote down 15 points that struck me as important which I will list in this and subsequent posts. They are not in any particular order nor very well digested but are here to share, and possibly discuss.
1.The intersection of the Social web and Semantic web is already happening, will be significant, and each will impact the other.
I start with an example, since it might help to explain some of the energy. FanHubz may at first site appear trivial, an apparent timewaster of the kind that gives Social Media a bad name. It’s a site which publishes data on programmes from the BBC and then allows people to form Fan clubs based around particular programmes, and then it selects and displays Tweets from people about that programme. So far so uninteresting, unless you happen to be a Doctor Who fan and a Twitter user in which case you are likely to be Tweeting while you watch. You could look for other people on Twitter but FanHubz makes it easier to find people like you who want to have a conversation while the programme is happening (or at any time). The site also gives you direct access to all that the BBC publishes about that programme. And it does that for all the BBC’s output, TV Radio and Music.
The building block for this work is that the BBC is now marking up all its content, segmented into usable chunks, using open protocols and standards. This means that the metadata about the content can be viewed on the web using a browser or, as in this case, used to dynamically update an application. These protocols are part of a common framework that allows data to be shared and reused across application, enterprise, and community boundaries.
One of the key players in that standard setting is a collaborative effort led by W3C with participation from a large number of researchers and industrial partners. It is based on the Resource Description Framework (RDF). The metadata is also accessible from FanHubz, as shown above, and there is a new Query Language for interrogating the metadata called SPARQL, also shown above. Talis is a UK company that is experimenting with the BBC and the UK Government, who are also working to make Government data available in open formats. Richard Wallis from Tallis demonstrated several fascinating experiments, listed here along with his slides.
2. A great deal of attention and effort is going into exposing, sharing, and connecting data using Universal Resource Indicators (URI) Linked data is a key stepping stone to a more wholly realised Semantic Web. It is more accurate and useful to talk about a Linked Data web at this stage than spend time debating, defining or speculating about The Semantic Web.
3. Searching for meaning is infinitely more complex and difficult than data integration.
There is enormous benefit in making data accessible through common metadata formats, and experimenting with what can be done by linking them. However, understanding what those linkages mean, and indeed being able to infer the many, many layers that make up the ‘meaning’ of a phenomenon or set of links needs to be approached using a range of separate tools, notably visualisation. Here is a demonstator visualising UK crime statistics within an ontology map
4. The focus of attention in a semantic web is on discovery and inference, not on search.
5. People understand and assign relevance, judge significance and associate emotion with items and elements. Machines can assess evidence. The intersection of the social and semantic web opens up interesting and significant opportunities for bridging those capabilities. One of themost often quoted features of the Social Web is that it forms an affective and trust-based filter on the web for its users. “If the news is important it’ll find me” was a quote to a US survey on political attitudes among young people, meaning that such people eschew traditional news outlets and rely on information from their trust network. Clearly this can causes problems, not least in the ‘tribalism’ observed in many Social Network Sites but it also means that the real-time stream of comment, status updates and tweets lays a trail for discovery, often with unexpected or random side-linkages, that is very different from the ordered world of a search engine with its mechanical algorithms based on formal relationships between web entities.
6. It’s happening anyway.
There are inevitably and properly questions to address to do with the relevance of all this to Development. However, people are already building schemas, prototypes, publishing data in open formats and developing some quite complex interconnected systems using semantic principles, such as in the already remarkably advanced Finnish Health Information Systems. Development agencies need to engage now, seek to understand the approach and what is happening, experiment to learn how to use the new tools and begin to envision how this can and will impact the people with whom we work. At its simplest, many Governments outside the OECD are simply not part of the influencing processes and bodies and will inherit an already mature set of standards and processes defined elsewhere. More interestingly, they manage and own huge data sets that inform increasingly sophisticated networks of linked data and need to develop themselves the skills and tools to be able to contribute and learn.
7. New value added services and service organisations will emerge
There are enormous opportunities for people and organisations to generate income from providing informational and knowledge services using the growing number of openly available data sets, combining them in novel ways and adding other tools that help discover meaning, facilitate transactions or simply have fun. The possibilites are well illustrated by simple applications now available for mobile phones, such as Yelp, that link location data to data and reviews about local businesses, restaurants and clubs. This final example is of a HyperLocal information resource, bringing together in one place openly available data on a particular community.
Filed under: Uncategorized