004 Datenverarbeitung; Informatik
Filtern
Dokumenttyp
- Dissertation (1)
- Bericht (1)
Sprache
- Englisch (2)
Volltext vorhanden
- ja (2)
Gehört zur Bibliographie
- nein (2)
Schlagworte
- Software Engineering (2) (entfernen)
Institut
- Wirtschaftswissenschaften (2) (entfernen)
Effort estimates are of utmost economic importance in software development projects. Estimates bridge the gap between managers and the invisible and almost artistic domain of developers. They give a means to managers to track and control projects. Consequently, numerous estimation approaches have been developed over the past decades, starting with Allan Albrecht's Function Point Analysis in the late 1970s. However, this work neither tries to develop just another estimation approach, nor focuses on improving accuracy of existing techniques. Instead of characterizing software development as a technological problem, this work understands software development as a sociological challenge. Consequently, this work focuses on the question, what happens when developers are confronted with estimates representing the major instrument of management control? Do estimates influence developers, or are they unaffected? Is it irrational to expect that developers start to communicate and discuss estimates, conform to them, work strategically, hide progress or delay? This study shows that it is inappropriate to assume an independency of estimated and actual development effort. A theory is developed and tested, that explains how developers and managers influence the relationship between estimated and actual development effort. The theory therefore elaborates the phenomenon of estimation fulfillment.
This paper documents the experiences of assurance evaluation during the early stage of a large software development project. This project researches, contracts and integrates privacy-respecting software to business environments. While assurance evaluation with ISO 15408 Common Criteria (CC) within the certification schemes is done after a system has been completed, our approach executes evaluation during the early phases of the software life cycle. The promise is to increase quality and to reduce testing and fault removal costs for later phases of the development process. First results from the still-ongoing project suggests that the Common Criteria can define a framework for assurance evaluation in ongoing development projects.