Abstract. Integrity constraints are incident part of conceptual models, including part of semantics of problem domain. Analysis of the most important methods of conceptual modelling has revealed that none of them analyze the complete set of integrity const
ISSN 1392 – 124X INFORMATION TECHNOLOGY AND CONTROL, 2005, Vol.34, No.4
REPRESENTATION OF INTEGRITY CONSTRAINTS
IN CONCEPTUAL MODELS1
Elita Miliauskait , Lina Nemurait
Kaunas University of Technology, Department of Information Systems
Studentu St. 50-308, LT-51368 Kaunas, Lithuania
Abstract. Integrity constraints are incident part of conceptual models, including part of semantics of problem do-main. Analysis of the most important methods of conceptual modelling has revealed that none of them analyze the complete set of integrity constraints needed for making semantically meaningful model. In our previous work the taxo-nomy of integrity constraints relevant for design of well-formed conceptual models was established. The goal of this paper is to extend capabilities of UML for required types of integrity constraints introducing stereotypes or reusing them from other methods. In contrast with current practice of deferring description of constraints to detailed design, modelling of constraints in the phase of conceptual analysis makes them reusable in various activities: not only in generating DB schema, but also in early verification, validation, transformation to other types of schemas and program code.
Keywords: information systems, integrity constraints, UML, OCL, profile, stereotypes, tags.
1. Indroduction
The raised level of abstraction of presentation of data and precision of modelling are gaining more and more meaning in nowadays. Growing volumes of data and possibilities of automatic dealing with data in e-Business, Data Warehouses, On-Line Analytical Pro-cessing, Data Mining and Semantic Web services require for more careful capturing of semantics of problem domain. OMG has defined Business Seman-tics of Business Rules 0 bringing the capability to express formally the meaning of expressions used in business (business vocabulary and business rules) independently of language. Conceptual model with derivation rules and integrity constraints serves as re-presentation of business terms at the conceptual level of information system, where transformations to logi-cal and physical representations are carried-out for processing and returning results to the business level. These transformations must be two-way and lossless, from business to physical level, as at current rate of business change, hand-written triggers or procedures untraceable from business are not acceptable for majo-rity of organizations. This is today’s vision of bridging business and information technologies.
The presentation of information on the logical and physical levels is well defined today in comparison with representation on conceptual and business level.
We have accepted 0 UML 0, 0 as the most suitable language for conceptual modelling. The ontological foundations for UML conceptual models are considered in 0, 0, where high-level stereotypes (kinds, subkinds, phases, roles, categories, mixins and role mixins) are proposed for well-founded conceptual modelling. Nevertheless, the precise definition of identities, relationships and other constraints essential for conceptual modelling, is not involved yet in re-lated UML research. In this paper, the possibilities to obtain precise conceptual model are of main impor-tance. Essentially, the problematic elements of con-ceptual models are integrity constraints that are often unfoundedly deferred to the phase of detail design. Such a situation is typical for existing not standard UML Data Modelling Profiles (e.g. 0), where only fundamental constraints are considered for logical data models on purpose of generation of database schema. Integrity constraints are incident part of conceptual models intensively used during analysis and design of information systems 0. They are essential for ensuring correctness of information model, and its ability to represent adequate semantics of problem domain. Integrity constraints may be implemented in data sto-rage system or become a part of software code, in order to protect against unallowable changes in data or invoking behaviour that may raise undesirable situa-tions in functioning information system.
1
The work is supported by Lithuanian State Science and Studies Foundation according to Eureka programme project IT-Europe” (Reg. No 3473)
Abstract. Integrity constraints are incident part of conceptual models, including part of semantics of problem domain. Analysis of the most important methods of conceptual modelling has revealed that none of them analyze the complete set of integrity const
E.Miliauskait , L.Nemurait
In conceptual model, integrity constraint is logical formula, dependent on problem domain, and it must be held true for all meaningful states of information system 0. Conceptual model with integrity constraints is semantically meaningful if constraints are consistent, and it is effective if constraints are not redundant. Inconsistent constraints raise errors and in-finite cycles in running software, and redundant const-raints worsen operation. Integrity constraints are consistent if system when performing transitions during its life cycle remains in consistent state. Cons-traints are redundant if they are overlapping or never can be violated. It is not possible to validate absolute consistency of integrity constraints, but it is possible to find and remove some kinds of inconsistencies. Therefore to create semantically meaningful con-ceptual model of good quality we should be able to capture required variety of integrity constraints and to validate them. During analysis of the most important methods of conceptual modelling (ER, EER, HERM, ORM, UML, xUML) the types of constraints that are important for well-formed conceptual models, and situations, when these constraints should be used, were established 0. The aim of this paper is to extend UML-related conceptual modelling method for capturing the whole important types of constraints. The design of data and behavioural schemas of va-rious forms (relational, object-relational, object-orien-ted, XML, or XML-based) of information system must be founded on the same conceptual model, and integrity constraints should be first class entities in it 0, 0.