Authors:
René Noël
;
Marcello Visconti
;
Gonzalo Valdés
and
Hernán Astudillo
Affiliation:
Universidad Técnica Federico Santa María, Chile
Keyword(s):
Extreme Programming, Evolutionary Design, Experimental Design.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence
;
Cross-Feeding between Data and Software Engineering
;
Evolutionary Design
;
Knowledge Engineering
;
Knowledge Engineering and Ontology Development
;
Knowledge-Based Systems
;
Software Engineering
;
Symbolic Systems
Abstract:
Evolutionary Design is Extreme Programming’s approach to organize software structure and its relationships, encouraging refactoring, test driven development and the simplest solution for the requirements of a single iteration, thus avoiding a big up-front design activity at the beginning of the project that can cause carrying on a huge structural complexity throughout the whole project. In order to contrast this approach with a planned or traditional design approach, an empirical evaluation of impact on software design quality and process productivity has been designed and conducted in an academic environment with toy size problems. Experimental studies planning details are presented, and two replications with different experimental designs are described. Results suggest that there are no differences in quality between both approaches, and that productivity is better when a planned design is adopted.