Adaptierbare CASE-Werkzeuge in prozeßorientierten Software-Entwicklungsumgebungen

Software development is a complex and demanding task, frequently carried out by teams of multiple developers. Software development methods define the steps to be taken and the products to be maintained in order to attain a successful project outcome. Thus, methods outline the structure of a project. In the early phases, upper-CASE tools are used to edit, check, and transform various types of software documents. The requirements for these tools vary considerably, depending on the organization, user role, task, and kind of system to
be developed. Consequently, tools have to be adaptable to the actual situation. Traditional CASE tools do not fulfill this demand sufficiently. Meta environments support tool developers in building customized tools. However, known systems are focused on only one aspect: The adaptation to different document types or the adaptation to different process models.
The thesis describes an approach for building process-centered software development environments (PSEEs). The resulting environments contain tools which are tailored to the given situation and are integrated with the software development process. An important component of these environments is an Object Management System (OMS). It serves as a central repository for both, software documents and information concerning the current process state. The OMS integrates the various tools with their different views on the common data-
base and offers numerous services which are used for implementing tool functions - among them views, access controls, transactions, and distributed notifications.
Thereby, the costs for building meta environments and for building their instances, i.e. concrete environments, are reduced. The services mentioned can only be used if data models are fine-grained and if tools implement an OMS-oriented architecture. In the thesis, this architecture has been implemented in a framework. It is used to build distributed multi-user tools. The framework consists of components for building CASE tools and process
tools. The configuration of tools depends on the data model of software documents and on additional information. These additional information are codified in tool parameters which
are associated to type definitions in the data model. The resulting tool specification is called
tool schema. Besides this black-box reuse, tool components can be extended with tool-specific
functionality. This white-box reuse requires a more extensive knowledge and offers more flexible means for tool adaptation.
For the description of process models we define a simple process modeling language which is used for demonstration purposes only. Based on the framework, a process engine can be
built. It controls process performance and can be flexibly extended with process-specific
functionality. CASE tools adapt themselves to the current process situation and allow their
users to influence process performance. Additionally, tools for planning and managing software processes have been built and have been integrated into the environment to underline the flexibility of the approach.