Realising the Flanders Research Information Space

Transcription

1 Realising the Flanders Research Information Space Peter Spyns & Geert Van Grootel published in Meersman R., Dillon T., Herrero P. et al., (Eds.): (eds.), Proceedings of the OTM 2011 Workshops, LNCS 7046, Springer Verlag, pp Flemish government, department EWI

2 Realising the Flanders Research Information Space Peter Spyns and Geert Van Grootel * Vlaamse overheid Departement Economie, Wetenschap en Innovatie Koning Albert II-laan 35, bus 10, B-1030 Brussel, Belgium {Peter.Spyns,Geert.VanGrootel}@ewi.vlaanderen.be Abstract. A model driven architecture combined with SBVR and fact oriented modelling are key standards and methodologies used to implement a Flemish research information government portal. The portal will in the future offer various services to researchers, research related organistions and policy makers. 1 Context The knowledge economy is one of the cornerstones of our modern society. Hence, it is very important for a country or region to have a good overview of its science and technology base to develop an appropriate policy mix of measures to support and stimulate research and innovation. Needing a new system for managing research information, The Flemish department of Economy, Science and Innovation (EWI) launched the Flanders Research Information Space programme (FRIS) [9]. The FRIS concept creates a virtual research information space covering all Flemish players in the field of economy, science and innovation. Within the space, research information can be stored and exchanged in a transparent and automated way. A first realisation is the FRIS research portal ( to provide up to date research information on projects, researchers and organisations in Flanders. One can browse for projects, individual researchers, research groups and collaborations. Also other European countries, which are member of the European Organisation for International Research Information 1, have deployed similar initiatives. A key feature is that data can be immediately collected at the point of creation in the operational processes of data providers (e.g., universities, funding bodies ). The data are up-to-date and are supposedly more accurate. Also, parallel data gathering processes can be eliminated, resulting in a lot of administrative work being spared. As many data providers are involved (universities, funding agencies, independent research institutes to only name the most important ones) it is not a surprise that the FRIS portal team is confronted with various (external) legacy systems, data heterogenity, different classifications, different storage formats, data of varying quality, different data input and upload techniques and so on. Having a standard E-R schema for the research domain, c.q. the CERIF model [3] 2, as interchange format is not enough to guarantee flawless interoperability and * In alphabetical order The Common European Research Information Format (CERIF) is a recommendation to EU member states for the storage and exchange of current research information. R. Meersman, T. Dillon, and P. Herrero (Eds.): OTM 2011 Workshops, LNCS 7046, pp , Springer-Verlag Berlin Heidelberg 2011

3 Realising the Flanders Research Information Space 139 information exchange with other (external) applications and services. Agreement (or least univocity) should be achieved on the terminology of the CERIF schema. 2 FRIS High Level Engineering Principles To realise a robust FRIS, the technical solution aimed at is to build a modelling process pipe line. Using a conceptual modelling method results in a collaboratively and agreed upon conceptual model that would automatically be converted in a CERIF-based E-R model that, in its turn, is implemented as optimised MySQL database tables. This corresponds to a model driven architecture. Mechanisms to transform business rules from a high level conceptual representation into efficiently executable low level software modules are needed and must be very robust to frequent updates of the business rules. Domain experts, not familiar with specific database modelling methods, should be able to adapt and extend the conceptual domain model without any further intervention on the lower levels. Therefore, EWI has decided to express the business facts in the domain concerned in an SBVR-style (Semantics of Business Vocabulary and Business Rules [5]). From the SVBR specification document [5: pp ] follows that fact oriented modelling methodologies e.g., Object Role Modelling (ORM) [2] can help in freeing domain experts from having to become familiarised with arcane and technical (at least to them) formalisms. Expressing the semantics of business rules in an easily understandable way for domain experts is paramount. Some even consider agreeing on the business rules more difficult than agreeing on the domain concepts and relationships [4]. 3 Two Cases 3.1 Business Vocabulary Definition To ground the vocabulary of the FRIS research portal, a concept definition storage is used. It provides access to the meaning of the concepts used by the FRIS. In the spirit of SBVR this is achieved by means of semantic communities (i.e. a group of people communicating on the same topic), which are subsequently divided into linguistic communities (according to the language the various subgroups of a semantic community use). The various communities can exchange meaning descriptions of concepts, map them on concept labels and determine rule sets. The SBVR standard of the Object Management Group (OMG) is said to be unique in its completeness and flexibility as it unifies several strong points: a combination of some ISO standards on terminology and vocabulary (e.g., ISO 37, 704, 1087); support for formal logics as basis via fact-oriented modelling; linguistics and linguistic annotation of grammar for natural language, including multilinguality; practical experience by application in large, complex organisations.

4 140 P. Spyns and G. Van Grootel In practice, the FRIS requirements imply not only semantically reconciliating several science classification schemes and lists of scientific topic codes but also more general ISO code tables on countries, languages, currencies as well as local Flemish standard etc. Hence, a list of concepts described as unambiguously as possible is required (together with a tool to manipulate and maintain this list). E.g., the science classification (IWETO) used by the Flemish fundamental research funding agency has been mapped into a classification of core technologies (based on the classification of Fraunhofer Institute of Systems and Innovation Research) used by the Flemish innovation funding agency. Thanks to this mapping industry patents can be related to research projects e.g. to implement a service that can determine the impact of research projects. To set up the semantic machinary behind the FRIS, EWI has contracted an external consultant 3 that has developped its own business semantics management methodology [1] and corresponding development environment based on a scientific fact oriented ontology engineering methodology and tools [6]. 3.2 Grounding Business Rules The above mentioned efforts address the problem of diverging interpretations of the business rules vocabulary [7,8]. A common conceptual model for the CERIF E-R schema and established a consensus amongst the involved parties (here: Flemish university data centres information analysts). Concepts of the semantic dictionary are combined to create semantic patterns (corresponding to binary fact types), which are then semantically restricted by constraints expressing specific business rules. These rules, i.e. specific constraints on the relationships, are kept separated from the semantics of the domain concepts and relationships. The CERIF concepts can be re-used. Its meaning stays the same even if the nature of the relation between the vocabulary changes in other rules for the same application (or even for other applications). 4 Discussion The stakeholders are positive regarding the approach and way of working as they recognise the advantages. In addition, the stakeholders appreciate the apparent promise on scalability and better maintainability of the data and knowledge in the FRIS research portal. Experiments have shown that the above described way of working can result in a substantial time and hence cost reduction [7]. However, many hurdles in particular data quality control at the input source still need to be overcome. 5 Future Work and Conclusion The FRIS portal is up and running for two years. In the future, the FRIS research portal will roll out a range of new services as part of the research information space. The possibilities are numerous: a white guide (who does what?), a library of publications by a particular researcher (digital library), a service for updating and reusing researchers CVs and provision of information on patents, to name but a few. It is 3 Collibra NV.

5 Realising the Flanders Research Information Space 141 expected that a model driven architecture and applying a fact oriented modelling methodology to define SBVR-style business rules and establish a semantic domain dictionary prove to be the right methods to realise these services. The cases presented in this paper illustrate two important steps towards this end. But it surely remains an ambitious, long term challenging endeavour and, for the Flemish government, it is an innovating 4 and fresh [ fris = fresh in Dutch] approach. References 1. De Leenheer, P., Christiaens, S., Meersman, R.: Business Semantics Management with DOGMA-MESS: a Case Study for Competency-centric HRM. Journal of Computers in Industry 61(8), (2010) 2. Halpin, T., Morgan, T.: Information Modeling and Relational Databases, 2nd edn. Morgan Kaufmann, San Francisco (2008) 3. Jörg, B.: The Common European Research Information Format Model (CERIF). Data Science Journal 9, (2010); Jeffery, K., Asserson, A. (eds.) 4. Meersman, R.: Ontologies and Databases: More than a Fleeting Resemblance. In: d Atri, A., Missikoff, M. (eds.) OES/SEO 2001 Rome Workshop. Luiss Publications (2001) 5. Object Management Group, Semantics of Business Vocabulary and Business Rules (SBVR) v1.0, 6. Spyns, P., Tang, Y., Meersman, R.: An ontology engineering methodology for DOGMA. Journal of Applied Ontology 1-2(3), (2008) 7. Spyns, P., Van Grootel, G., Jörg, B., Christiaens, S.: Realising a Flemish govern-ment innovation information portal with Business Semantics Management. In: Proceedings of CRIS 2010, pp (2010) 8. Spyns, P., Van Grootel, G.: Re-engineering Business Rules for a Government Innovation Information Portal. In: Meersman, R., Dillon, T., Herrero, P. (eds.) OTM LNCS, vol. 6428, pp Springer, Heidelberg (2010) 9. Van Grootel, G., Spyns, P., Christiaens, S., Jörg, B.: Business Semantics Management Supports Government Innovation Information Portal. In: Meersman, R., Herrero, P., Dillon, T. (eds.) OTM 2009 Workshops. LNCS, vol. 5872, pp Springer, Heidelberg (2009) 4 In 2009, the FRIS received the innovation prize by the Flemish Government.