In addition to 'extended model' and 'multi-model' approaches, we envisage a need for enhanced traceability between the conceptual behavioral models and more 'high level' models used at earlier stages of the IS development process (e.g. models used in the F3-approach). Our assumption here is that as the complexity of models increases, so will the variation in expectations and needs between users. To address this difficulty, we anticipate a meta-model approach (repository) in order to handle this complexity, something which we are currently addressing in the context of two (of the total of five) different 'high-level' conceptual models with preserved traceability between the Enterprise Model level and the conceptual model.
We believe that, as a complement to communication over various schemes, various techniques for browsing, visualization, animation etc. (see e.g. Bubenko and Wangler (1992), Lundell (1997)) using populated models, might significantly improve end-users' understanding of behavioral models, something which we currently also address in terms of a tool for external validation of conceptual schemes represented in a repository based on CDIF.
'A schema is expressive when it represents requirements in a natural way and can be easily understood through the meaning of ER schema constructs, without the need for further explanation.' (p. 141)Langefors' early work (Langefors 1973) constitutes a foundation for "the Scandinavian Approach" to Information Systems. As part of his early work, he defined the infological equation (Langefors 1973), based on the assumption that information I is a function not only dependent on data, but also on pre-knowledge S and available time t for interpretation (Langefors 1995 p. 144) of the observer:
For example, traditional ER Diagrams implicitly capture some behavioral aspects, in terms of navigational semantics for update propagation, see e.g. Dogac and Chen (1983):
'Although the Entity-Relationship Diagram is a non-directed graph, when the existence dependency information is considered, a directed graph is obtained. This graph is then used to guide update propagation.' (p. 361)Hence, there is a potential conflict if an explicit (conceptual) rule language is introduced to be used in conjunction with ER Models, with implicit semantics expressed in ER Diagrams.
Our current position is that the implicit semantics of ER Diagrams ought to be extracted and kept in a common conceptual 'rule-base,' thereby facilitating one (conceptually) coherent storage of conceptual rules. Related to this issue, current work at the University of Skövde addresses extensibility aspects of CDIF with respect to conceptual business rules and their transformation to SQL3.
'The ER model has been most successful as a tool for communication between the designer and end user during the requirements-analysis and conceptual-design phases because of its simplicity, ease of understanding, and convenience of representation.' (p. 167)Contrary to this, Moody (1996) claims that 'the understandability of Entity Relationship models to users is largely a myth and is unsubstantiated in practice' (p. 241).
Experiences from using the F3 methodology, which includes 'informal' conceptual diagrammatic notations for external validation purposes, are promising (Persson 1997). Bubenko (1993) stresses 'that verification and validation of informal models must be done informally by a user-customer-architect dialogue and "walk-through" of the models (descriptions) so far produced' (p. 18).
The question remains, how does one assess expressiveness of these models when there is an inherent subjectivity, in the form of S, in the infological equation?
Based on a Qualitative approach, Lundell and Lings (1997) present an analysis of, and suggest a methodological complement to the ISO standard for CASE-tool evaluation (ISO 1995). One aspect of this methodology includes a complex process in which the relevant set of criteria for a specific organization must be identified. In that paper we suggest a Qualitative approach to this evolving process aimed at establishing an 'optimal' set of characteristics for a specific requirement within the organization.
Similar to this CASE-tool evaluation, we envisage that a qualitative approach would be appropriate as one approach for identifying relevant characteristics for diagrammatic notations to be used in behavioral models.
To us, it is not apparent that there today exists one specific graphical notation that would be most suitable for describing behavior. However, certainly such models (or parts of specific models) which are used in various IS methodologies (e.g. F3's activities and usage model) could be included as resources in the data collection phase in qualitative analysis, when organizations need to identify appropriate notations suitable to their own needs. In order for such an 'organizational notation' for behavioral models to be effectively used, there would be a need for adaptable CASE-tools.
We consider diagrammatic notations used in this process, and the related external validation of developed diagrams, as a human communication process. Hence, it is a social process that involves end-users and other stakeholders. Consequently, if one adheres to this view, then the conceptual modeling community can gain from exploring methodologies originated within the social sciences such as, for example, Grounded Theory.
Bubenko, Janis A. jr., and Wangler, Benkt (1992) Research Directions in Conceptual Specification Development, SYSLAB, Dept. of Computer and Systems Science, Stockholm University, and SISU - the Swedish Institute for Systems Development, Sweden, Report No. 91-024-DSV, Nov. 1991.
Bubenko, Janis jr. (1993) Extending the Scope of Information Modelling, Invited paper: 4th International Workshop on Deductive Approach to Information Systems and Databases, Lloret, Costa Brava, Sept. 20-22 1993.
Dogac, Asuman and Chen, Peter Pin-Shan (1983) Entity-Relationship Model in the ANSI/SPARC Framework In Chen, P. P. (ed.), Entity-Relationship Approach to Information Modeling and Analysis, Elsevier Science Publ., North-Holland, pp. 357-374.
ISO (1995) Information Technology - Guideline for the evaluation and selection of CASE tools, ISO/IEC JTC1/SC7/WG4, ISO/IEC 14102:1995(E).
Langefors, Börje (1973) Theoretical Analysis of Information Systems (4th ed.), Studentlitteratur, Lund, Sweden, AUERBACH Publishers Inc., Philadelphia.
Langefors, Börje (1995) Essays on Infology: Summing up and Planning for the Future, Edited by Bo Dahlbom, Studentlitteratur, Lund.
Lundell, Björn (1997) Penning a Methodology: A Classroom example used as a framework for reasoning about Information Systems Development, The British Computer Society: Specialist Group on Information Systems Methodologies, 5th Annual Conference on Methodologies: Training and Education of Methodology Practitioners and Researchers, Preston UK, 27-29 Aug. 1997 (accepted for publication).
Lundell, Björn and Lings, Brian (1997) Evaluation and selection of CASE-tools within the ISO framework: qualitative issues, The Eighth Australasian Conference on Information Systems (ACIS '97) , Adelaide, Australia, 29th Sept.-2nd Oct. 1997 (accepted for publication).
Moody, Daniel (1996), Graphical Entity Relationship Models: Towards a More User Understandable Representation of Data, in Thalheim B. (ed.) Conceptual Modeling - ER '96: 15th International Conference on the Entity-Relationship Approach, Cottbus, Germany, Oct. 1996, pp. 227-244.
Persson, Anne (1997) Using the F3 Enterprise Model for Specification of Requirements - An Initial Experience Report, The Second CAiSE97/IFIP8.1 International Workshop on Evaluation of Modeling Methods in Systems Analysis and Design, Barcelona, June 1997.
Tanaka, Asterio K. (1992) On Conceptual design of Active Databases, PhD thesis, Georgia Institute of Technology, December 1992.
Yang, Dongqing, Teorey, Toby J. and Fry, James P. (1987) A Practical Approach to Transforming Extended ER Diagrams into the Relational Model, Information Sciences, Vol. 42, pp. 167-186.