| Rocky road to interoperable Europe |
|
|
| 11 April 2006 | |
|
Semantic interoperability in Europe is being realised one small step at a time, writes Stephen Gardner. The goal of interoperability is an ambitious one, involving vast and complex challenges. But, in technical terms at least, many European Union (EU) member states have made significant progress. For semantic interoperability however, the path ahead remains long and tortuous. Government departments and agencies want to exchange data in meaningful form, and this requires common indexes, catalogues and lexicons. In cases where local authorities store data, cataloguing standards can vary considerably even by comparison with the authority next door. Magnify this up to regional and national levels and the huge task of determining common vocabularies becomes apparent. On top of all this, the EU wants to encourage interoperability between its member states, so citizens and businesses can more easily take advantage of their right to move across borders. Transferrable pension rights are a practical example – Germans retiring to Spain, for example, will want their benefits to be paid on time after processing through an efficient, interoperable system. The European Commission's Interchange of Data between Administrations (IDA) programme has been the prime mover in setting the standards to enable this. In 2004, IDA published the European Interoperability Framework – a set of policies, guidelines and standards for interoperability. In 2005, IDA, now re-christened IDABC (Interoperable Delivery of European eGovernment Services to Administrations, Businesses and Citizens), followed this up with a Content Interoperability Strategy (CIS). This defines semantic interoperability as one of three main elements of interoperability, the others being organisational and technical. “Technical interoperability is covered well with concepts, standards and protocols,” says Professor Herbert Kubicek of the Bremen, Germany based Institute for Information Management. “But semantic interoperability is much more complex. Some areas are more advanced because there is a tradition of common standards, such as in the case of centralised vehicle registration systems. But in countries where local authorities control car registration, they need to find common agreements.” One case of a tradition of interoperability, says Professor Kubicek, is the road traffic accident compensation recovery system in the UK, whereby the costs of NHS treatment of accident victims can be charged to insurance companies via the Department of Work and Pensions' Compensation Recovery Unit (CRU). There is interoperability, including common semantic standards, between the systems of the CRU and the insurers. This is grounded in legislation and therefore all the elements are in place: technical, legal, organisational and semantic. “In general, small and centralised countries such as Austria and Denmark are quite advanced,” says Professor Kubicek. “But more important than the size of the country is how many standards exist. Semantic interoperability on a European level is very far away.” Dreaming of common standards Speaking at a European Commission event in Brussels in September 2005, Hosein Askari from Denmark's National IT and Telecoms Agency explained that the Danish approach was driven by “a dream of common standards.” Denmark's Ministry of Finance established a committee to chase the dream in 2000. This committee ensured the implementation of XML as the public sector communication standard, bringing schemas together into a clearinghouse known as InfoStructureBase. Most recently this has published naming and design rules for creating XML schemas in compliance with national standards. Peter Brown, Senior Expert at the Austrian Federal Chancellory, explained at the same workshop that Austria has pursued a similar approach. An interoperable communications architecture has been created, addressing organisational and technical issues as well as semantic. Working groups at different levels of government are centrally coordinated and open standards are considered a prerequisite. “The issue is coordination and knowing who does what,” commented Brown. “There needs to be a governance and sustainability model to overcome limits of project-based work.” Meet MICHAEL The sectors where most progress has been made in applying interoperable architectures for the purposes of data mining are, perhaps unsurprisingly, libraries and archives. The European Commission has funded a number of projects to explore the possibilities in these areas. One such is the MICHAEL project – Multilingual Inventory of Cultural Heritage in Europe. By 2007, MICHAEL will have set up a multilingual platform allowing users a single point of access to Europe's national cultural portals. It builds on a Franco-Italian 'archive of archives' project which describes digital collections held by libraries, archives, museums and other cultural institutions in both countries and allows searching of the top-level catalogue. A project along the same lines is the Network of Excellence on Digital Libraries, known as DELOS. This has the ambitious aim to define a comprehensive theory for 'the life-cycle of Digital Library information', and includes a 'knowledge extraction and semantic interoperability' research theme. Coordinating this is Liz Lyon, Director of the UK Office for Library Networking at the University of Bath. She reinforces Professor Kubicek's point about some areas being more ready than others for semantic interoperability because of 'tradition'. Libraries and the cultural heritage sector apply semantic interoperability “by habit,” she says. “There is a historical precedent for describing things. Some disciplines have formal thesaurae and core ontologies that underpin the main knowledge-based systems.” Interoperability and privacy
In a more general sense, however, the limits of semantic interoperability have yet to be addressed, especially in relation to electronic identity (eID) schemes, which most European governments are developing. From a convenience point of view, administrations would be very happy if eID data could be shared and mined across the whole gamut of government services. |
| < Prev | Next > |
|---|
