Middleware in automation serves the connection of hardware and software components of applications in automation-like control or human-machine interfaces. To cope with today's flexibility requirements, there is an urgent demand to not only communicate data but to describe the semantics of the distinct elements of the distributed information space. Therefore, middleware in automation needs to provide some means for information modeling. To derive general concepts and recommendations on how to implement information models on middleware in automation, we discuss several relevant standards that provide partial information models. We then identify the fundamental information modeling constructs of these partial models and derive general requirements for the information modeling capabilities of middleware in automation.

Nowadays the complexity of design processes, no matter which design domain (CAD, software engineering, etc.) they belong to, requires system support by means of so-called repositories. Repositories help managing design artifacts by offering adequate storage and manipulation services. Some of the most important features of a repository are version management and activity management. Versioning comprises the specification, storage, and maintenance of versioned design objects whereas activity management is responsible for cooperation control, designflow management and management of design transactions processing versioned design objects. Regarding these issues (version and activity management) repository technology, as we think, should not only provide pre-defined services, but should be flexible enough to reflect different application needs. For that reason, we propose to provide repository managers by generic methods, i.e. by generating the corresponding functionality. In this paper, we consider a representative cooperation model, which is based on versioning services, in order to identify the major data manipulation and activity control needs of cooperative design applications. We will focus on the data manipulation needs by introducing our generative approach for customizing versioning facilities. Additionally, we will outline our ideas of applying a generative approach also for the provision of tailored activity control services. Thus, the paper wants to show that by exploiting generic methods and reuse as well as the extensibility properties of new object-relational database technology, repository managers can be flexibly tailored to special application needs and, thereby, applications do not have to be forced to deal with systems only providing pre-defined services.