Search

Modern enterprise engineering (EE) requires deep understanding of organizations and their interaction with their context. Because of this, in early phases of the EE process, enterprise context models are often built and used to reason about organizational needs with respects to actors in their context and vice versa. However, far from simple, this task is usually cumbersome because of knowledge and communication gaps among technical personnel performing EE activities and their administrative counterparts. In this paper, we propose the use of strategic patterns expressed with the i* language aimed to help bridging this gap. Patterns emerged from several industrial applications of our DHARMA method, and synthesize knowledge about common enterprise strategies, e.g. CRM. Patterns have been constructed based on the well-known Porter’s model of the 5 market forces and built upon i* strategic dependency models. In this way technical and administrative knowledge and skills are synthesized in a commonly agreeable framework. The use of patterns is illustrated with an industrial example in the telecom field.

When an organization faces the need of integrating some workflow-related activities in its information system, it becomes necessary to have at hand some well-defined informational model to be used as a framework for determining the selection criteria onto which the requirements of the organization can be mapped. Some proposals exist that provide such a framework, remarkably the WfMC reference model, but they are designed to be appl icable when workflow tools are selected independently from other software, and departing from a set of well-known requirements. Often this is not the case: workflow facilities are needed as a part of the procurement of a larger, composite information syste m and therefore the general goals of the system have to be analyzed, assigned to its individual components and further detailed. We propose in this paper the MULTSEC method in charge of analyzing the initial goals of the system, determining the types of components that form the system architecture, building quality models for each type and then mapping the goals into detailed requirements which can be measured using quality criteria. We develop in some detail the quality model (compliant with the ISO/IEC 9126-1 quality standard) for the workflow type of tools; we show how the quality model can be used to refine and clarify the requirements in order to guarantee a highly reliable selection result; and we use it to evaluate two particular workflow solutions a- ailable in the market (kept anonymous in the paper). We develop our proposal using a particular selection experience we have recently been involved in, namely the procurement of a document management subsystem to be integrated in an academic data management information system for our university.

The definition of the context of a system is one of the most relevant activities in the early phases of information systems engineering. It allows system engineers to narrow the system scope, by defining well established system boundaries. In practice, outlining a system context model is complex and cumbersome. In order to support context modeling, in this paper we propose a catalogue of context model elements expressed in i*, which can be reused as building blocks in the construction of context models for new systems. We describe the process used for the identification of a set of actors and dependencies recurrently appearing in several academic and industrial cases, and the process to store them into a catalogue of reusable i* context dependencies.

In this paper we propose a taxonomy for classifying COTS business applications, i.e. products that are used in the daily functioning of all types of organizations worldwide, such as ERP systems and document management tools. We propose the identification of characterization attributes to arrange the domains which these products belong to, and also we group these domains into categories. We define questions and answers as a means for browsing the taxonomy during COTS selection. We show the need of identifying and recording the relationships among the domains and propose the use of actor-oriented models for expressing these relationships as dependencies. Last, we explore the definition of quality models for the domains, to be used in COTS selection, focusing on their reusability and stepwise definition downwards the hierarchy. (Best Paper Award 2004)