Claudia Pons

A Lightweight Approach for the Semantic Validation of Model Refinements

Model Driven Engineering proposes the use of models at different levels of abstraction. Step by step validation of model refinements is necessary to guarantee the correctness of the final product with respect to its initial models. But, given that accurate validation activities require the application of formal modeling languages with a complex syntax and semantics and need to use complex formal analysis tools, they are rarely used in practice. In this article we describe a lightweight validation approach that does not require the use of third-party (formal) languages. The approach makes use of the standard OCL as the only visible formalism, so that refinements can be checked by using tools that are fully understood by the MDE community. Additionally, for the efficient evaluation of the refinement conditions a hybrid strategy that combines model checking, testing and theorem proving is implemented. Correctness and complexity of the proposal are empirically validated by means of the development of case studies and a comparison with the Alloy analyzer.

Foundations of Object-Oriented Modeling Notations in a Dynamic Logic Framework

The Unified Modeling Language UML is a language for specifying object-oriented systems. It is a standard in the domain of object-oriented analysis and design. Due to the missing formal foundation of the UML the syntax and the semantics of a number of UML costructs are not precisely defined.

Dependency Relations between Models in the Unified Process

The goal of the Unified Process is to guide developers in efficiently implementing and deploying systems that meet customer needs. During the Unified Process, a variety of models of the system are developed. All these models are not independent, they are related to each other. Elements in one model have trace dependencies to other models: they are semantically overlapping and together represent the system as a whole. It is necessary to have a precise definition of the syntax and semantics of the different models and their relationships, since the lack of accuracy in their definition can lead to wrong model interpretations and inconsistency between models. We distinguish three different kinds of dependency relations between models and propose a formal description of them. The goal of the proposed formalization is to provide formal foundations for tools that perform intelligent analysis on models expressed in UML assisting software engineers throughout the development process.

Formal foundations of object-oriented modeling notations

Describes and classifies the different solutions that have been proposed to realize the integration of graphic modeling languages, which are known and accepted by software developers, with formal modeling languages having analysis and verification tools. Inspired by that classification, we define a new integration proposal, based on first-order dynamic logic. The principal benefits of the proposed formalization can be summarized as follows. (i) The different views of a system are integrated in a single formal model; this allows us to define compatibility rules between the separate views, on both a syntactic and a semantic level. (ii) Using formal manipulation, it is possible to deduce further knowledge from the specification. (iii) Faults in the specifications, expressed using a user-friendly notation, can be revealed using analysis and verification techniques based on the formal kernel model. The principal difference between this model and other object-oriented formal models is that it integrates both of the levels in the modeling notation architecture into a single conceptual framework. The integration of modeling entities and modeled entities into a single formalism allows us to express both static and dynamic aspects of either the model or the modeled system within a first-order formalism.

PAMPERO: precise assistant for the modeling process in an environment with refinement orientation

Abstraction [2] facilitates the understanding of complex systems by dealing with the major issues before getting involved in the detail. Apart from enabling for complexity management, the inverse of abstraction, refinement, captures the essential relationship between specification and implementation. Refinement relationship makes it possible to understand how each business goal relates to each system requirement and how each requirement relates to each facet of the design and ultimately to each line of the code. Documenting the refinement relationship between these layers allows developers to verify whether the code meets its specification or not, trace the impact of changes in the business goals and execute test assertions written in terms of abstract model’s vocabulary by translating them to the concrete model’s vocabulary.

A Domain Specific Language for the Development of Collaborative Systems

Domain-specific languages (DSLs) are high level languages defined for combining expressivity and simplicity by means of constructs which are close to the problem domain and distant from the intricacies of underlying software implementation constraints. This paper presents a language to graphically document the analysis and design decisions embodied in collaborative system development. The language was designed as a conservative extension of the UML and it enables the application of the MDD approach to the development of such systems.

Design method for a Historical Data Warehouse, explicit valid time in multidimensional models

We present a new storage structure that contains the explicit valid time called Historical DW. This proposal combines a Historical Data Base and a DW in one integrated model. The objective of this model is to solve the temporal limitations of the traditional multidimensional structures. Although the Temporal DW considers, besides the temporal dimension, other aspects related to time, this model only takes into account the changes that occur in the DW schema, both in dimensions and in hierarchies. Therefore, considering the need for registering values that allow to evaluate trends, variations, maximum and minimum values, a problem to be solved is how to shape the values of entities, attributes or relationships that may vary in time, in the design of the structure. The fact that the needed data was stored is already known, but the temporal search mechanisms would be complex. The Historical DW design is framed in the MDD approach, a strategy that allows, through successive automatic transformations, to obtain one of the implementations of the model in a Relational Data Base Management System (RDBMS).

Graphical Query Mechanism for Historical Data Warehouse within MDD

Query tools that depend on the ability of programmers, impose a cognitive load that could reduce the users productivity. Within MDD, the proposal of our work is the creation of a visual query mechanism derived from a historical multidimensional data structure. This mechanism facilitates (and partially automates) the formulation of temporal and decision making queries.

A Two-Level Calculus for Composing Hybrid QVT Transformations

The standard for model transformations QVT offers two dialects: Relations Language and Operational Mappings Language. Each one of these dialects can be used in isolation, resulting in purely declarative transformations or purely imperative transformation respectively; alternatively, both dialects can be combined resulting in a hybrid transformation approach. On the other hand, the availability of compositional approaches to produce complex transformations from smaller units is a major concern in the area of model transformations. Compositional approaches for pure QVT transformations are supported by a number of tools; however no composition technique exists that can consistently manage the hybrid approach. Such partial techniques provide suitable answers to most practical needs; but they do not cover the entire composition spectrum. The aim of this article is to describe a technique for composing model transformations embracing both dimensions - declarative and imperative - so that the hybrid approach can be smoothly supported. Additionally, we report the implementation of a software tool supporting such technique and we sketch its validation.

Practical Verification Strategy for Refinement Conditions in UML Models

This paper presents an automatic and simple method for creating refinement condition for UML models. Conditions are fully written in OCL, making it unnecessary the application of mathematical languages which are in general hardly accepted to software engineers. Besides, considering that the state space where OCL conditions are evaluated might be too large (or even infinite), the strategy of micromodels is applied in order to reduce the search space. The overall contribution is to propitiate the performing of verification activities during the model-driven development process.