Marcela Ridao

Domain independent regularities in scenarios

Scenario is a description technique which has attracted much attention not only from practitioners but also from researchers. Literature on this topic shows the possibilities that this description technique provides to enhance the understanding of task-related descriptions and the communication among stakeholders. On the other hand, the idea of pattern, as a description of a problem which occurs over and over again in our environment and its solution, has also attracted lots of attention, especially in software design. The paper describes an initial taxonomy of domain-independent-scenario patterns and proposes a preliminary process to use this taxonomy for scenario construction.

Defining System Context Using Scenarios

It is already well known that in order to define the software system requirements deep knowledge of the universe of discourse is needed. This includes not only what is actually happening but also what the plans for the future are. Unless the requirements engineer has a previous knowledge of the domain, he or she may become overloaded by data of different nature. To cope with this, a gradual process which slowly migrates from learning to definition is advisable. During the learning phase, the focus should be placed on the current facts which should be organized and recorded. Later, the forthcoming plans should be also understood but instead of recording them as they are, they should be used to design future universe of discourse models. This chapter focuses on the description of current situations and lays foundations for the description of future situations by means of an evolution process.

An ATL transformation from natural language requirements models to business models of a MDA project

MDA is a software development framework where the core is a set of automatic transformations of models. One of these models, the CIM (Computer Independent Model), is used to define the business process model. Though a complete automatic construction of the CIM is not possible, we have proposed the integration of some natural language requirements models and we have defined a strategy to derive a CIM from these models. In this paper, we present an ATL transformation that implements this strategy to obtain a UML class diagram representing a preliminary CIM from requirements models. This transformation fits with MDA approach.

User selectable reduced instruction set forth

Abstract This article describes the construction of a Forth cross-compiler whose main characteristic is the resources use optimisation, based on the knowledge of the users application. This optimisation consists in the elimination of portions of the language not required for the functioning of the developed application. The number of Forth words available for elimination by the post-optimiser was notoriously increased due to the change introduced in the execution model as a consequence of the migration from an interpreter-compiler to a cross-compiler.