Baltic IT&T

ICT Events



ICT Market




Baltic IT&T 2011

ICT Calendar


Hosted by:

About journalSite Map
Home > Articles > E-government


The Architecture of Semantic Interoperability in Estonias State Registries

Martin Luts, Certified Information System Auditor, Estonia

The Department of State Information Systems of the Estonian Ministry of Economic Affairs and Communications has just released a document that is called Instructions and Methodology for the Semantic Description of Databases and Operations Performed by Databases.


Interoperability is defined by the Institute of Electrical and Electronics Engineers (IEEE) as the ability of two or more systems or components to exchange information and to make use of the information that has been exchanged.[1]  The European Interoperability Framework defines semantic interoperability as one of the three elements of interoperability as such.  The others are organisational and technical interoperability.[2]  Semantic interoperability refers to the ability of organisations to understand exchanged data in the same way.  The semantic interoperability of information systems is their ability to make adequate use of data received from other software systems.  All three dimensions of interoperability organisational, technical and semantic are considered to be prerequisites for the seamless and user-centred delivery of pan-European E-government services.
Such services are built upon the exchange of information, and in future, they will be technologically based on state-of-the-art XML technologies.  Semantic or content interoperability refers to the need to ensure that the meaning of exchanged information is not lost in the process, that it is readable and understandable for the people, applications and institutions that are involved.  Before cross-border services can be delivered, European institutions and member states are going to have to agree on a host of semantic specifications such as descriptions of people, products, processes, forms, etc.  There will also have to be agreement on how best to formulate these descriptions and on where they should be stored for public use.[3]


There are several arguments in favour of semantic interoperability:
1) The quality of data improves, data from different sources can be integrated, and there are fewer errors and inaccuracies when using the data and making decisions (this relates to the misinterpretation of data, as well as discrepancies which arise from a duplication of data);
2) Less investment in the production and acquisition of data is needed, since it is easier to reuse existing data;
3) Participants have to spend less time in integration the information systems of various organisations;
4) Sustainability of the application or information system increases, and it is easier to ensure further developments.  The knowledge base that is related to information systems is preserved even after key personnel have left the organisation;
5) The level of necessity for data can be established, and retention periods can be determined.
In brief, organisations are able to save money in the production of data if they integrate external data sources.  Integration, of course, is not free of charge reuse of data has its price, too.  Semantic descriptions (records) must be maintained, different data structures must be integrated, and there is the need to resolve conflicts which arise from semantic diversity (e.g., terms with different fields of meaning).

In order to improve semantic interoperability, Web service descriptions are enriched with semantic information which makes reference to semantic resources.


Let us assume that one local government defines the word street as a hard-surface road with sidewalks, while another defines it as a road of any kind of surface and with sidewalks.  The issue is allocation of funds for the winter maintenance of roads, and this depends on the surface area.  The different definitions offered by the two local governments will clearly mean difficulties in ensuring the proper allocation of funds.


Semantic interoperability refers not just to the exchange of data, but also to the exchange of metadata among organisations.  Semantic interoperability involves an exchange of information about the context of data, their relations, operations and overall functioning.  Semantics determine the way in which the elements of the data structures which are exchanged relate to real world objects, relations and events.  An appropriate way to achieve semantic interoperability is to describe the exchanged data explicitly and in detail, i.e., to define the data.
Semantic description can occur at different levels of formality:
Readable by humans describing operations in a way which makes it possible for information system developers (architects, analysts, designers, software designers, etc.) to locate the needed operation, to establish its suitability, and to integrate it with their own applications;
Processed by machines defining operations in a formal language (RDF(s),  OWL, etc.) so that it is at least partly possible to identify and integrate operations in an automatic way.


The Department of State Information Systems has just released two new documents Instructions for the Semantic Description of Databases and Operations Performed by Databases,[4] and Methodology for the Semantic Interoperability of Databases and Operations Performed by Databases.[5]  The two documents are also available in English.[6]  The European Semantic/Content Interoperability strategy[7] is quite general in terms, while Instructions for the Semantic Description provide more concrete guidelines and rules for application owners, developers and auditors.


Semantic interoperability, which we now understand to mean the ability to understand exchanged data in a similar way and to make adequate use of the data, requires a system for the semantic description of data in the databases and the input and output data that are used in performing operations (e.g., Web services).  The architecture for such a system is based on the following principles:
1) The principle of reuse a single data element or operation should be described only once, and in future uses, the initial description must be accepted;
2) Compilation, management and search for the descriptions of data elements and operations must be simple and standardised.
The architecture of semantic interoperability (Figure 1) consists of the following components:
1) The domain glossary a central description of the semantics of data elements.  The description is divided up into domains such as land survey.  Languages used in the semantic description of the domain glossary are based on the W3C recommendations known as RDF (Resource Description Framework), RDFS (RDF Schema) and OWL (Web Ontology Language).[8]  The description is stored in the administration system of the state information system (RIHA).  What does this mean for developers of information systems?  Best practice at this time involves the creation of a domain glossary as part of documentation as the information system is developed.  This is done in the form of a table within the specification.  Developers then have to compile the glossary in another format RDFS/OWL, and it can later be transformed into a glossary which can be read by individuals (e.g., transformation of XSL into HTML, plain text, etc.).
2) Semantic descriptions of databases and operations.  This refers to a description of every single database or operation, as well as its components.  If possible, there must also be reference to the respective entry in the domain glossary.  Languages for description are WSDL and SA-WSDL (Semantic Annotations for WSDL and XML Schema)[9]  The idea of the SA-WSDL recommendation is simple it standardises the way in which Web service descriptions are semantically enriched by adding pointers to WSDL.  These refer to the semantic description of a Web service element in a model outside the WSDL document.  The model could have an OWL ontology or an UML model that has been created on the basis of, for example, the Ontology UML Profile (OUP).[10]  An example of this would be a description of the cadastral unit data element in a Web service.  Instead of describing it in human-readable form in a WDSL document, we would use an SA-WSDL pointer which refers to the appropriate record in land survey ontology.
One thing to consider is whether to store metadata within the information resource (WSDL, for instance, which includes semantic description) or outside of it.  The semantic description of a database field is located in the domain glossary, and the described resource includes a link to the metadata entry.  Our suggestion is that operations be described in a WSDL file with the respective database, but also that the data be stored centrally for easier access.  The terms of a domain should always be described centrally.  Descriptions of databases (data structures) and the input/output elements of operations must include references to a central glossary.  If descriptions are stored outside of the information resource, that requires more work creation and maintenance of repositories, as well as supervision of those who are preparing the descriptions.  Of course, this approach might mean that describers are more highly motivated, because there would be more supervisors in the sense that the descriptions would be available to everyone.


Here is the main way to enhance semantic interoperability between organisations or information systems.  First, describe the exchanged data and performed operations semantically.  In other words, create terms and definitions for these data by focusing on their meaning.  Then name the data structures and operations in a way which reveals their nature or function as explicitly as possible.
The primary goal in achieving semantic interoperability between state registers in Estonia is to create a human-readable glossary with examples for the input and output data of the system, as well as a description of the operations.  An adequate description is needed before proceeding to the next level, and it must refer to domain and inter-domain ontology and multilingualism, as well as to machine processes automatic identification, launch and composition of the operation.
Pilot projects have been launched at the Estonian Land Cadastre and the Population Registers.  In future issues of our Journal we hope to report on how the semantic operability architecture is working.

Authors contacts: martin.luts(at)

[1]   The IEEE Standard Computer Dictionary: A Compilation of IEEE Standard Computer Glossaries (1990).
[2]   See
[3]   See
[4]   See
[5]   See
[6]   See
[7]   See
[8]   See
[9]   Vallner, U.  Nationwide Components of Estonias State Information System, Baltic IT&T Review, No. 3(42), 2006.
[10]   See

Rating - 10 (5 Votes)
My rating:
My comment:
Web site owners do not carry responsibility for materials submitted by commentators and reserve rights to delete ones violating legal regulations and breaching of the decencies. Thank you for understanding!


Top news
Instantly online - 17 golden rules to combat online risks and for safer surfing mobile social networks [3]
Experts Warn of ICT Labour Shortage and Loss of Competitive Edge in Europe by 2015
Lithuanian organisations has been active to participate in CIP ICT PSP third call

Question for readers
How secure you evaluate your information and data?
Highly secure
Quite secure, but some security improvements are needed
Security is insufficient
No security at all