Abstract
Software estimation is a vital yet challenging project management activity. Various methods, from empirical to algorithmic, have been developed to fit different development contexts, from plan-driven to agile. Recently, machine learning techniques have shown potential in this realm but are still underexplored, especially for individual task estimation. We investigate the use of machine learning techniques in predicting task effort and duration in software projects to assess their applicability and effectiveness in production environments, identify the best-performing algorithms, and pinpoint key input variables (features) for predictions. We conducted experiments with datasets of various sizes and structures exported from three project management tools used by partner companies. For each dataset, we trained regression models for predicting the effort and duration of individual tasks using eight machine learning algorithms. The models were validated using k-fold cross-validation and evaluated with several metrics. Ensemble algorithms like Random Forest, Extra Trees Regressor, and XGBoost consistently outperformed non-ensemble ones across the three datasets. However, the estimation accuracy and feature importance varied significantly across datasets, with a Mean Magnitude of Relative Error (MMRE) ranging from 0.11 to 9.45 across the datasets and target variables. Nevertheless, even in the worst-performing dataset, effort estimates aggregated to the project level showed good accuracy, with MMRE = 0.23. Machine learning algorithms, especially ensemble ones, seem to be a viable option for estimating the effort and duration of individual tasks in software projects. However, the quality of the estimates and the relevant features may depend largely on the characteristics of the available datasets and underlying projects. Nevertheless, even when the accuracy of individual estimates is poor, the aggregated estimates at the project level may present a good accuracy due to error compensation.

Abstract

Abstract

Abstract

Abstract
Adaptive Object Models, though a well-known architectural pattern, is seldomly used in software projects where, due to their nature, would highly benefit from it. Characteristics such as complexity, reduced literature and case-studies, lack of reusable framework components, and fundamental issues as those regarding runtime evolution, drive developers away. By overcoming these barriers with a set of patterns, tools and applications, and addressing pending research problems, Adaptive Object Models can dramatically alter the way developers design their software. This paper presents a survey in the field, describes the preliminary contributions and outlines the ongoing doctoral work.

Abstract
Algebraic specification languages have been successfully used for the formal specification of abstract data types (ADTs) and software components, and there are several approaches to automatically derive test cases that check the conformity between the implementation and the algebraic specification of a software component. However, existing approaches do not assure the coverage of conditional axioms and conditions embedded in complex axioms. In this paper, we present a novel approach and a tool to automatically derive test cases from bounded algebraic specifications of ADTs, assuring axiom coverage and of all minterms in its full disjunctive normal form (FDNF). The algebraic specification is first translated into the Alloy modelling language, and the Alloy Analyzer tool is used to find model instances for each test goal (axiom and minterm to cover), from which test cases in JUnit are extracted.