Joost Noppen

Software development with imperfect information

Delivering software systems that fulfill all requirements of the stakeholders is very difficult, if not at all impossible. We consider the problem of coping with imperfect information, like interpreting incomplete requirement specifications or vagueness in decisions, one of the main reasons that makes software design difficult. We define a method for tracing design decisions under imperfect information. To model and compare requirements with estimations, we present fuzzy and stochastic techniques. This approach offers adequate decision support that can deal with imperfect information during software design. The approach is illustrated by a real-world example, based on a storm surge barrier system.

Imperfect requirements in software development

Requirement Specifications are very difficult to define. Due to lack of information and differences in interpretation, software engineers are faced with the necessity to redesign and iterate. This imperfection in software requirement specifications is commonly addressed by incremental design. In this paper, we advocate an approach where the imperfect requirements in requirement specifications are modeled by fuzzy sets. By supporting this approach with a requirement tracing and an optimization approach, the necessity for design iteration can be reduced.

Unanticipated Software Evolution

This workshop was dedicated to research towards better support for unanticipated software evolution (USE) in development tools, programming languages, component models and related runtime infrastructures. The report gives an overview of the submitted papers and summarizes the essence of discussions during plenary sessions and in working groups.

Market-driven approach based on Markov decision theory for optimal use of resources in software development

Changes in requirements may have a severe impact on development processes. For example, if requirements change during the course of a software development activity, it may be necessary to reschedule development activities so that the new requirements can be addressed in a timely manner. Unfortunately, current software development methods do not provide explicit means to adapt development processes with respect to changes in requirements. The paper proposes a method based on Markov decision theory, which determines the estimated optimal development schedule with respect to probabilistic product demands and resource constraints. This method is supported by a tool and applied to an industrial case.

Dealing with fuzzy information in software design methods

Software design methods incorporate a large set of heuristic rules that should result in stable software architecture of high quality. In general, clearly defined inputs are required to deliver the desired results. Unfortunately, especially in the early phases of software development, it is very difficult or even impossible to provide precisely defined information. Since methods cannot deal with imprecision, the designers need to make approximations which are generally not justifiable. In this paper, we will advocate an approach where the inputs for software design methods are modeled by using fuzzy sets. This approach renders the need for introduction of extra information for removal of inexactness obsolete.

Dealing with imprecise quality factors in software design

During the design of a software system impreciseness can manifest itself in for instance the requirements or performance estimations. While it is common to eliminate the impreciseness by information that can not be justified, it is better to model the impreciseness since it is the most accurate description that is available at the current point in time. In this paper we present an approach, which allows the explicit specification of quality estimations and quality requirements including the imprecise nature. In this approach the impreciseness is modeled and addressed using representations from probability theory and fuzzy set theory.

Imperfect information in software product line engineering

In this chapter, we examine the phenomenon of imperfect information, the problems it causes during SPL engineering and we outline a generalised approach for addressing these problems. In the final section of this chapter we will examine the way forward for achieving life-cycle wide supprt for imperfect information in SPL engineering.

Comparison of two approaches to dynamic programming

Both in mathematics and in computer science Dynamic Programming is a well known concept. It is an algorithmic technique, which can be used to write efficient algorithms, based on the avoidance of multiple executions of identical subcomputations. Its definition in both disciplines is however quite different. The aim of this paper is to compare both definitions. It is shown that the computer science approach is more efficient, since it avoids the execution of unneeded subcomputations, whereas the mathematical approach has greater possibilities of reducing memory requirements.