Abstract. Today's researchers argue that facilitating interoperability between applications by relying on agreements concerning the format and meaning (i.e. syntax and semantics) of exchanged data and the ordering of these exchanges are not enough to achieve a complete, effective and meaningful collaboration. The use of data (i.e. pragmatics) is important as well. Pragmatic interoperability thus requires that the use of data be mutually understood between collaborating systems. However, we observe that the notion of pragmatic interoperability is still largely unsettled, as evidenced by the various proposed definitions and the lack of a canonical understanding. Therefore, our objective is to contribute to a more in-depth understanding of this concept through a systematic review of published definitions. Our results show that, indeed, various interpretations of pragmatic interoperability exist. If we categorize the concepts that can be derived from these definitions, we see two broad categories: system level and business level. Within the scope of each of these levels, we see some degree of agreement among the definitions. However, comparing the definitions across these levels, we observe no general agreement. At the system level, a shared understanding of the intended and actual use of exchanged system message in a given context is essential. At the business level, pragmatic interoperability goes beyond service use by considering also the compatibility of business intentions, business rules, organizational policies, and the establishment and maintenance of trust and reputation mechanisms between collaborating business parties.

In hybrid service composition approaches, processes are used to describe the core part of the composition logic whereas the rules are used to specify decision making constraints and conditions. These rules are exposed as services and are used in the processes whenever a certain decision has to be made. To evaluate these rules, we need a mechanism 1) to share data between the main process and the rule service; 2) to decide upon which service to invoke based on the result received from the rule service. In this paper, we propose a tuple space based solution for supporting these requirements. We also include descriptions of an application scenario to motivate our work, a prototype of the proposed solution and we demonstrate how our solution helps in achieving process flexibility with minimal maintenance costs in the context of changing requirements.

Driven by dynamic market demands, enterprises are continuously exploring collaborations with others to add value to their services and seize new market opportunities. Achieving enterprise collaboration is facilitated by Enterprise Application Integration and Business-to-Business approaches that employ architectural paradigms like Service Oriented Architecture and incorporate technological advancements in networking and computing. However, flexibility remains a major challenge related to enterprise collaboration. How can changes in demands and opportunities be reflected in collaboration solutions with minimum time and effort and with maximum reuse of existing applications? This paper proposes an approach towards a more flexible integration of enterprise applications in the context of service mediation. We achieve this by combining goal-based, model-driven and serviceoriented approaches. In particular, we pay special attention to the separation of business rules from the business process of the integration solution. Specifying the requirements as goal models, we separate those parts which are more likely to evolve over time in terms of business rules. These business rules are then made executable by exposing them as Web services and incorporating them into the design of the business process. Thus, should the business rules change, the business process remains unaffected. Finally, this paper also provides an evaluation of the flexibility of our solution in relation to the current work in business process flexibility research.

Pragmatic interoperability (PI) is the compatibility between the intended versus the actual effect of message exchange. This paper advances PI as a new research agenda within the gamut of enterprise interoperability research. PI is timely in today's new enterprises as it is increasingly important that organizations are able to collectively add value to their products and services through effective collaboration. When enterprise systems exchange information, PI goes beyond the compatibility between the structure and the meaning of shared information. It also considers the use of information in a given context as an equally important aspect. As PI is a relatively new domain, this paper attempts to elucidate its notion by identifying and relating its key concepts from proposed definitions, and by reviewing extant approaches to identify critical knowledge gaps in PI research.

During the course of our participation in the Semantic Web Services (SWS) Challenge, we have shown how the concepts defined in the COnceptual Services MOdeling (COSMO) framework for the modeling, reasoning and analysis of services can be used to solve the Mediation Problem Scenarios of the Challenge. Along with the service-oriented refinement and composition paradigm of COSMO, our approach is also based on model-driven and goal-oriented principles where the semantic integration of applications is designed at a layer of abstraction higher than technology specifications. The objective of this paper is to evaluate our previous and current research efforts towards advancing our solution to the semantic integration of service-oriented applications, particularly, using the mediation problem scenarios of the Challenge. We do this by presenting the state of the art of our solution while reporting our experience with applying our solution to the scenarios including lessons learned and identified research challenges.