The origin of the integration of structured data does not date from yesterday. The idea of a structured web goes back to 1994 with the creation of the W3C led by Tim Berners-Lee, father of the web, confirmed in 1998 (with the beginning of the work in 1999). It will take a few years of work to define the first protocols and formats.
A rise in industrial power
From 2004 onwards, several protocols and formats to structure data appeared: RDF (Resource Description Framework, the basic language of the Semantic Web), its complement RDF Schema (which brings together processes and tools to define the ontologies that structure RDF resources); OWL (Web Ontology Language, which defines the RDF vocabulary) and SPARQL (a language based on RDF to query data resources).
All these tools, although they allow data to be structured, have yet to be widely used in information management tools. The major groups are not mistaken and “we have seen an increase in industrial use since 2007,” says Fabien Gandon, a researcher at Inria (the French National Institute for Research in Computer Science and Control).
For him, six main issues are at the heart of the development of Web 3.0. First, the confirmation of standards, which benefit from the feedback of 6 years of experience, and redefined in the RDFa 1.1. Then, the massive implementation of the Semantic Web through large companies (Oracle, IBM but also Yahoo even if the question arises since the integration of its search engine in Microsoft Bing, or Google which seems to do “websem” without admitting it). “The full deployment requires tools that are ready,” says the researcher.
Creating an ecosystem around data
Putting public data online is also among the most important scenarios to contribute to the diffusion of the technology. But the issue is necessarily political and its dissemination is likely to be delayed as the interests of some (citizens in particular) are not necessarily those of industry (which wants to monetize this value of information).
“We have a lot of trouble getting people to understand that we can create an ecosystem around data and not value on the real time processing of information”, Nicolas Chauvat regrets. Another issue is skills: “One out of every two phone calls I receive is about a profile search,” says Fabien Gandon. There is a real demand for semantic web engineers and technicians, and also a need for decision makers. In other words, as long as the critical mass of skills is not reached, there is no salvation for the semantic web.
An obstacle that the propagation of the dynamics to go beyond the implementation of standards will allow to reach. This obviously implies continuing research work, particularly through the scaling of processing, taking into account new uses such as mobility and social networks, but also the quality of data and the need to keep interaction as simple as possible through interfaces adapted to democratize the use of the semantic web.
The future of companies faced with mountains of data is at stake, which tends to become more common after a certain size. “Semantic web technologies are emerging as a solution for information management,” concludes the Inria spokesperson.