Practical Experience with the First Fully-Integrated Bachelor Programme Based on the Knowledge Standard SBVR
On December 11, 2007, the standards organization OMG declared SBVR (Semantics of Business Vocabulary and Business Rules) to be a standard. The name of the standard seems to suggest that it is primarily concerned with 'business' in a commercial, for-profit sense. The truth is that this standard also opens up new possibilities in education and IT, as well as in many other fields. In education, SBVR gives the opportunity to drop the prevailing stovepipe approach and implement a completely integrated approach. This new approach is characterised by full integration of all subjects. The authors followed such a bachelor education with the CogNIAM (short for Cognition enhanced NIAM) speech community , a best business practice of SBVR. In this paper we share our experiences with this new generation of education.
1 Statistics Netherlands
Statistics Netherlands is responsible for collecting, processing, and publishing statistics to be used in practice by policymakers and for scientific research. In addition to its responsibility for (official) national statistics, Statistics Netherlands also has the task of producing the Dutch part of the statistics for the European Community.
2 Business case for upgrading employees
Within Statistics Netherlands it is clear — from a historical point of view — that quite a few legacy systems are still in place, an inheritance from the industrial age. In the year 2000, the Business Statistics (BES) division, by far the largest division of Statistics Netherlands, introduced the concept of treating at least twenty-five stovepipe systems in a generic way. This was a major step forward. It clearly demonstrated the resulting productivity increase. Statistics Netherlands wanted to build on this tangible success.
Mid-2003, the decision was made at Statistics Netherlands to give approximately forty employees substantial aid in attaining their bachelor's degrees in knowledge management. This challenge was offered to PNA Training, which offers a fully integrated educational approach based on fact orientation. A major aim of this upgrading process was to capture the tacit knowledge of these employees in such a way that this knowledge could be consulted through a structured database. As the structured concept definitions are all recorded in natural language, there is easy comprehension.
An important part of this bachelor programme was to secure knowledge relevant to Statistics Netherlands through practical courses and a final thesis. Furthermore, during the course of the project the student-employees were integrally coached on their knowledge as well as their competencies and attitude. All courses were taught in complete interrelation to one another, based on what is called in SBVR the generic component of the conceptual schema, or the meta-model of fact orientation.
This was new in two ways. First of all, knowledge, competencies, and a positive attitude were integrated. Secondly, there was integration across the various courses offered. In the current marketplace this integral approach is increasingly sought in an employee. This empowers an employee with competencies that he can use during his entire life, yielding positive consequences such as flexibility and comprehensive understanding.
Accordingly, the chosen curriculum did not consist of individual courses provided by individual teachers. In principle there was only one teacher who initially taught all the courses and also facilitated the practical subjects and coached the student during his bachelor thesis. In this way the student-employee was prepared for an integrated professional career in a knowledge-intensive environment.
Most student-employees captured certain business process descriptions during their practical assignments and thesis work. Concept definitions and fact type forms — crucial elements of SBVR — were hardly ever supplied in these traditional use cases. Instead, textual descriptions were usually supplied, which quite often contained information that was (1) incomplete, (2) unaccountable, (3) inconsistent, or (4) superfluous.
Over the last three years, considerable experience has been gained at Statistics Netherlands, mainly within the division of Business Statistics where the focus has been question articulation for statistical investigation and explicit requirements specification. Within Statistics Netherlands as a whole, the authors have probably gained the most practical experience with question articulation and explicit requirements specification. We wish to share this experience and continue to achieve high-quality results as an example of a better IT approach.
3 Experiences with fully-integrated education based on fact orientation and SBVR
Before the fully-integrated bachelor programme, we had received all our training in the well-known stovepipe system of traditional education. Usually this implied learning a subject by heart, taking an exam, and subsequently often forgetting the information. It is now obvious that these subjects were all beautiful but temporary 'islands'. The first thing we learned using the new integral approach was that all knowledge and competencies were required to be mastered by the student not only at exam time but at any subsequent moment in time as well.
This educational approach was also used in our bachelor’s education: the competencies were gradually acquired in a manner that started with the available knowledge network of the learner, until the approach was fully assimilated by the student. Moreover, as an outcome of the practical assignments we completed, we discovered that the fact-based approach could be useful not only for the IT world but for the world of the business end-user as well.
Describing a subject (or problem) using the methodology can be quite challenging at first but, as is well-known, practice makes perfect. As a matter of fact, in our daily work we currently try to define every subject in this way. This implies making the distinction between structurally relevant and irrelevant issues and producing clear and understandable concept definitions. In the end, the tacit knowledge that resided in the heads of many employees of Statistics Netherlands could be structurally expressed by means of SBVR. Using this approach we are able to secure the knowledge in SBVR format, ensuring that the knowledge was not lost when an employee left the company.
The widespread practical use of the structured knowledge standard SBVR can have many positive consequences.   Of course, there will be enormous pressure from conservative forces to block this transition from the industrial stovepipes to the knowledge economy integration. It is worthwhile mentioning the sentence attributed to Thomas Watson, Chairman of IBM, in 1943: "I think there is a world market for maybe five computers."
Complete interrelationship at a concrete level
One aspect of all subjects in this curriculum was the consistent use of the CogNIAM knowledge triangle. By understanding the deep structure of knowledge for which part of the triangle is a blueprint, we were able to experience how specifications could be developed much more productively. In general, one could say that most end-users do fairly well with their concrete daily work; however, most end-users have difficulties with abstract formulations. We have learned that both concrete illustrations and abstract representations are necessary, but it is important that each is used for its proper purpose.
4 Results of applying SBVR
In this section we give a quantitative summary of the two bachelor theses. This entails numbers regarding concept definitions, fact types, fact type forms, constraints, and derivation rules (the last two are called rules in SBVR).
4.1 The heart of the ABR
The ABR ("Algemeen Bedrijven Register" — Dutch Statistical Business Register) is one of the most important information systems for Statistics Netherlands. The system supports the set-up of statistics (among others, the statistics regarding business productivity and national accounts). At the onset there was no clear definition of the interrelationships between the various collections of the ABR. There was too little insight into the structure of the collections and the interrelationships of the collections. Furthermore, the concept definitions needed to be formulated in a clear and comprehensive way. The understandability of the ABR's structure could be enhanced by defining a number of comprehensible examples in a common language that could be understood by everyone.
First of all, the existing processes (AS-IS) were mapped, using mainly the core of the ABR. In addition, the concepts were defined in a clear manner, taking care to eliminate circular references. Last but not least, several recommendations were made with respect to the HABR (TO-BE) — the 6th generation of the ABR. Of course the fact type diagrams were accompanied by sentence patterns (known as 'fact type forms' in SBVR) and business rules.
The end result was a clear enrichment to Statistics Netherlands — in particular, the self-explanatory structure and comprehensible examples. The thesis is consulted on a regular basis through the intranet of Statistics Netherlands.
Concept definitions 756 Roles (variables) 422 Fact type diagrams 56 Fact type forms 254 Constraints 198 Derivation rules 15
4.2 The CDC (Confrontation Data Concerns)
The significance of the large companies populating the TOP250 companies of Statistics Netherlands is crucial with respect to the business statistics of Statistics Netherlands, thereby implying that correctness, timeliness, and consistency are of utmost importance. After all, the data quality of the leading companies, the TOP250, to a large extent defines the quality of the branch data published by Statistics Netherlands. The quality of the data can be insured by promptly noticing possible inconsistencies, by developing the desired unified view of the statistical data, and by taking swift actions to correct potential problems. In order to ensure the data quality for the TOP250 companies, it is necessary to store the data, which has been brought together by several statistical departments, in one place and check for plausibility and consistency.
The problem was formulated as follows:
Specify and design an automated system for the concern coordinators where the various available and consultable sources within Statistics Netherlands are combined in order to be able to perform the desired unified view. This unified view has to be possible for at least all TOP250 companies.
An inventory was made of the requirements from the sector manager and the users of the CDC (Confrontation Data Concerns — the name given to the prototype geared to solving the problem described above). To this end, our procedure relied heavily on the use of concrete valid and invalid examples. The necessary statistical sources have been mapped. A complete and integral list of concept definitions has been composed, through cooperation with the stakeholders. The fact type diagrams and the accompanying sentence patterns (i.e., fact type forms in SBVR), as well as the business rules, have been described in a clear and comprehensive way.
The definition of the CDC has led to the development of a prototype. This prototype is a generic application using an interface that has been fine-tuned in cooperation with the user. Furthermore, a clear SBVR-based IT-description is now available. The experiences reported by the users have been positive; after all, an application has been developed in which the demands and wishes of the users have been taken seriously and are clear and recognizable. The prototype will be developed further in the near future by the IT department of Statistics Netherlands.
Concept definitions 801 Roles (variables) 264 Fact type diagrams 68 Fact type forms 267 Constraints 268 Derivation rules 38
We have acquired various competencies during our fact orientation-based, fully-integrated bachelor study, using the CogNIAM approach — a best business practice of SBVR. An example is the structured way of working with knowledge. In addition, we have gained a better insight into the world of the business user and the IT experts, and their association.
We have learned to aim for total coherence wherever possible. We believe that this should be part of all education and have therefore recommended the introduction of this approach throughout the entire Statistics Netherlands organization. Our experience with this approach is that it appears to be very demanding in the beginning. Once the good results come in, it becomes second nature to professionally analyze nearly every knowledge aspect  in this way.
We would recommend introducing SBVR (e.g. by utilizing the business practice CogNIAM) on a much larger scale in high quality institutions such as Statistics Netherlands. We conclude with the famous quote of André Rieu, in Toronto, December 2007: "Every person has the possibility to perform a job he likes in his life." Being solidly trained in meta-cognitive competencies makes this a reality.
 Semantics of Business Vocabulary and Business Rules (SBVR), v1.0, Object Management Group (Jan. 2008). Available as document 08-01-02 at http://www.omg.org/spec/SBVR/1.0/PDF. SBVR 1.0 and supporting files are available at http://www.omg.org/spec/SBVR/1.0/
 CogNIAM, Cognition enhanced NIAM, the methodology and product suite supporting the business practice of NIAM, URL: http://www.pna-group.nl
 G.M.A. Verheijen and J. Bekkum, "NIAM: an information analysis method," Information Systems Design Methodologies: a comparative review, Proc. IFIP WG8.1 Working Conf., Noordwijkerhout, The Netherlands, North Holland Publishing (1982), pp. 537-590.
 According to SBVR, the conceptual schema includes a generic component and a domain-specific component. The generic component is common to all conceptual schemas — for example, relevant axioms from logic and mathematics. The domain-specific component includes the concept definitions and declarations of the ground fact types and business rules relevant to the specific business domain. See SBVR v1.0, p. 89.
 PNA Training obtained a renewed accreditation for this bachelor programme in 2006 from the Accreditation Organization of the Netherlands and Flanders, the highest accreditation organization in the Netherlands and Flanders for higher education. NVAO, Accreditation Organization of the Netherlands and Flanders, URL: http://www.nvao.nl
 'Question articulation for statistical investigation' involves the process of composing well-formed questions for specific statistical subject areas. This concerns the various variables (or roles) used in the subject and clear concept descriptions of these variables in natural language. Strong use is made of comprehensible, elucidating examples.
 Sjir Nijssen, "SBVR: Semantics for Business," Business Rules Journal, Vol. 8, No. 10 (Oct. 2007), URL: http://www.BRCommunity.com/a2007/b367.html
 Sjir Nijssen, "A Process to Specify the Most Important Business Rule in SBVR," Business Rules Journal, Vol. 8, No. 12 (Dec. 2007), URL: http://www.BRCommunity.com/a2007/b380.html
 Ronald G. Ross, "How Rules and Processes Relate ~ Part 6. Point-of-Knowledge Architecture (POKA)," Business Rule Journal, Vol. 7, No. 3 (Mar. 2006), URL: http://www.BRCommunity.com/a2006/b276.html
# # #