Semantic interoperability and integration is concerned with the use of explicit semantic descriptions to facilitate information and systems integration. Due to the widespread importance of integration, many disparate communities have tackled this problem. They have developed a wide variety of overlapping but complementary technologies and approaches. The seminar has the following objectives:

1. To stimulate collaboration between diverse communities bound by common objectives in the area of semantic interoperability and integration;
2. To lay the foundation for a framework and a theory for understanding and classifying technologies for semantic interoperability and integration;
3. To set the research agenda for this research area the long-term aim of building a "research pipeline" for creating and disseminating results in industry.

Highlights of the Week

Feature talks, short talks, and panel discussions were structured roughly around four main themes: Mapping and translation, industrial experiences, theoretical foundations, and standards and benchmarks. Different days were devoted to different themes.

1. Mapping and Translation
2. Industrial Experiences
3. Theoretical Foundations
4. Standards and Benchmarks

Breakout Sessions

The concluding session on Friday morning was devoted mainly to report back on the discussion and outcomes of the various breakout sessions and in establishing future actions in the field of semantic interoperability and integration.

1. Social Aspects
2. Use Cases and Requirements
3. Mapping Typology and Tools
4. Mapping Notations and Languages
5. Theoretical Foundations
6. Infrastructure and Architectures

Conclusion

We set out to achieve three main objectives in the area of semantic integration and interoperability: 1) to stimulate collaboration, 2) to lay the groundwork for a future comprehensive framework for understanding the field and 3) to set a research agenda.

We successfully brought together experts from industry, academia and government representing historically separate communities including: database integration, category theory, standards, digital libraries, ontologies, knowledge representation, and the semantic Web. The varied program including invited talks,shorter talks, panels and breakout sessions provided the context for much collaboration. We achieved this objective.

Much of the week was spent wrestling with differences in terminology, identifying a range of relevant broad issues, puzzling over more specific and often subtle distinctions that arise in different sub-areas. The topics for the breakout sessions, form a good starting point for the eventual development of a more comprehensive framework that will provide a background for understanding and comparing different techniques, tools and applications that are developed in coming years.

Finally, we considered a range of issues that need to be included on agenda for future research. The problem of semantic interoperability and integration is hard and it is not clear how it could be solved in the near future. For instance, we don’t know how to formally specify the problem yet, let alone solving it. Hence it is necessary to work on two fronts: theoreticians need to make their case for the appropriate foundations upon which semantic interoperability and integration can be formalized, while practitioners and users need to expose their local semantics for the benefit of knowledge sharing.

The Semantic Web provides a playground for experimentation, but it also introduces new problems. There are a lot of challenging infrastructure issues still to be addressed and standardization efforts are still at their infancy. Furthermore there is a lack of lengthy experiences and large-scale scenarios to evaluate the scalability of current methods and techniques.

The discussions and outcomes of talks, breakout sessions and panels during the seminar have highlighted these issues and helped to put together efforts which were previously conducted separately in different communities. More discussion and cross-disciplinary collaboration is needed, but the first steps in converging and reaching a consensus might already be well under way.