Methodical support for model-driven software engineering with enterprise models

A central research goal in information systems science is to achieve a close alignment between business processes, structures of organizations, and the functionality offered by enterprise information systems (EISs), which are used to support the work of organizations. Traditionally, there is a methodical gap between describing organizational incidents and software functionality, because organizations and software systems are understood and constructed with different terminology and on different levels of abstractions, typically also by differently educated groups of people. In enterprise models (EMs), dedicated modeling language elements are used to express knowledge about processes in organizations, e. g., about who is responsible for performing actions, what resources are involved, and what strategic goals are intended to be realized by organizational means. The work at hand shows, how EISs can be created based on this knowledge, which serve as supporting software for performing these tasks. Software development traditionally has to face a distinction between people who work with software, and people who create software. With the use of EMs, a chance opens up to closer involve the users of software systems into the process of developing and configuring the software. Building software from enterprise models is desirable, because once a dedicated relationship between enterprise models and software functionality has been established by a development method, involved users and responsible stakeholders can adapt the software according to their business needs, without having to deal with programming or technical details. This increases efficiency both in developing and operating the software, because software functionality is derived from requirements implicitly stated in EMs. Such a development procedure also promises to more efficiently adapt EIS to dynamic changes in organizations and their environment. Trust in the developed software system is also increased by involving users and responsible stakeholders into specifying the resulting software functionality. The following research work elaborates a software development method to create EISs from EMs. The method is designed as a generic framework to work with any enterprise modeling language, and to generate software for any target system platforms, after appropriate configuration. Fundamental challenges in methodically transforming conceptual models to implementation artifacts, are faced by involving auxiliary models into the software creation process, and splitting up the transformation procedure into multiple dedicated phases. Using this approach, the abstraction gap between conceptual enterprise models and technical implementation artifacts gets systematically bridged by introduced methodical concepts, in order to perform an ontological turn from the bird’s-eye-view description perspective of enterprise models, to an internal system perspective describing technical details of a software system. The elaborated method provides means for efficiently guiding modelers and software developers through the software engineering process. It can be configured at multiple points, to choose the degree of automation on a continuum between a manually supervised development process with methodically scheduled manual development steps, and a zero-coding 100% code generation approach. To clarify the theoretically introduced concepts, prototypical implementation examples are included in the present work. They demonstrate how to configure the method with model-transformations, validity checks, and domain-specific modeling languages, and serve as initial example cases for enterprise model driven software development using the Software Engineering with Enterprise Models (SEEM) method.