Alternative representations for visual constrainst specification in the layered view model

Abstract

Extensible Markup Language (XML), with its rich set of semantics and constraints, is becoming the dominant standard for storing, describing and interchanging data among various Enterprises Information Systems (EIS) and databases. With the increased reliance on such semi-structured data and schemas, there exists a requirement to model, design, and constrain semi-structured data and the associated semantics at a higher level of abstraction than at the instance or data level. But most semi-structured schema languages lack the ability to provide higher levels of abstraction, such as visual constraints, that are easily understood by humans. Conversely, though Object-Oriented (OO) conceptual models offers the power in describing and modelling real-world data semantics, constraints and their inter-relationships in a form that is precise and comprehensible to users, they provide insufficient modelling constructs for utilizing XML schema like data descriptions and constraints. Therefore, it is interesting to investigate conceptual and schema formalisms as a means of providing higher level semantics in the context of XML-related data engineering. In this paper, we present a visual constraint specification model for an XML layered view model. First we briefly outline the view model and then provide a detailed discussion on modelling issues related to view constraint specification using two OO modelling languages, namely OMG's UML/OCL and XML Semantics (XSemantic) nets. To demonstrate our concepts, we also provide an illustrative case study example based on a real-world application.