Michael Halling

Predictive Regressions with Time-Varying Coefficients

We evaluate predictive regressions that explicitly consider the time-variation of coefficients in a comprehensive Bayesian framework. For monthly returns of the S&P 500 index, we demonstrate statistical as well as economic evidence of out-of-sample predictability: relative to an investor using the historic mean, an investor using our methodology could have earned consistently positive utility gains (between 1.8% and 5.8% per year over different time periods). We also find that predictive models with constant coefficients are dominated by models with time-varying coefficients. Finally, we show a strong link between out-of-sample predictability and the business cycle.

Investigating the defect detection effectiveness and cost benefit of nominal inspection teams

Inspection is an effective but also expensive quality assurance activity to find defects early during software development. The defect detection process, team size, and staff hours invested can have a considerable impact on the defect detection effectiveness and cost-benefit of an inspection. In this paper, we use empirical data and a probabilistic model to estimate this impact for nominal (noncommunicating) inspection teams in an experiment context. Further, the analysis investigates how cutting off the inspection after a certain time frame would influence inspection performance. Main findings of the investigation are: 1) Using combinations of different reading techniques in a team is considerably more effective than using the best single technique only (regardless of the observed level of effort). 2) For optimizing the inspection performance, determining the optimal process mix in a team is more important than adding an inspector (above a certain team size) in our model. 3) A high level of defect detection effectiveness is much more costly to achieve than a moderate level since the average cost for the defects found by the inspector last added to a team increases more than linearly with growing effort investment. The work provides an initial baseline of inspection performance with regard to process diversity and effort in inspection teams. We encourage further studies on the topic of time usage with defect detection techniques and its effect on inspection effectiveness in a variety of inspection contexts to support inspection planning with limited resources.

Using reading techniques to focus inspection performance

Software inspection is a quality assurance method to detect defects early during the software development process. For inspection planning there are defect detection techniques, so-called reading techniques, which let the inspection planner focus the effectiveness of individual inspectors on specific sets of defects. For realistic planning it is important to use empirically evaluated defect detection techniques. We report on the replication of a large-scale experiment in an academic environment. The experiment evaluated the effectiveness of defect detection for inspectors who use a checklist or focused scenarios on individual and team level. A main finding of the experiments is that the teams were effective to find defects: In both experiments the inspection teams found on average more than 70% of the defects in the product. The checklist consistently was overall somewhat more effective on individual level, while the scenarios traded overall defect detection effectiveness for much better effectiveness regarding their target focus, in our case specific parts of the documents. Another main result of the study is that scenario-based reading techniques can be used in inspection planning to focus individual performance without significant loss of effectiveness on team level.

Integrating Collaborative Processes and Quality Assurance Techniques: Experiences from Requirements Negotiation

Collaboration is essential in many mission-critical activities. Consequently, numerous methods and tools are available supporting collaborative processes such as strategic planning, risk management, requirements definition, and so on. These methods typically emphasize the collaborative, value-creating activities, but there is often less emphasis on quality aspects. Quality assurance (QA) techniques have been wellknown in engineering for a long time, and their effectiveness and efficiency has been empirically evaluated in many domains. In this paper, we propose to integrate repeatable QA techniques and collaborative processes. We evaluate our idea in the context of a collaborative process for requirements negotiation. We propose pre-process techniques to be used before the actual negotiation, in-process techniques for checking quality during a negotiation, as well as post-process inspection techniques. These techniques help a project team reduce unnecessary complexity and to mitigate risks stemming from defects in requirements negotiation results. We present the results of a feasibility study we conducted to test our approach.

An economic approach for improving requirements negotiation models with inspection

Stakeholder goals identified during requirements elicitation are usually informal and incomplete statements about a system considered for development. There are numerous approaches for capturing such informal models. For example, we have found the EasyWinWin requirements negotiation method to be an efficient way for attaining consensus among the success-critical stakeholders. The WinWin negotiation model captures stakeholder goals as win conditions, issues, options and agreements. When such a model has to be transformed into more formal representations, quality becomes particularly important. Approaches for validating such informal models can increase quality and provide guidance for further refinement of requirements. Inspection is a proven approach to identify defects and is also applicable to early life cycle artifacts. This paper reports on an empirical study demonstrating the usefulness of an inspection technique for requirements negotiation models. The study employs a conservative economic model, which considers the effect of defect slippage during development on defect detection benefits from inspection. The main finding of the study is that inspection is an economic validation technique for requirements negotiation models. There are, however, certain limitations that need to be studied in more detail.

Investigating the influence of inspector capability factors with four inspection techniques on inspection performance

We report on a controlled experiment with over 170 student subjects to investigate the influence of inspection process, i.e., the defect detection technique applied, and inspector capability factors on the effectiveness and efficiency of inspections on individual and team level. The inspector capability factors include measures on the inspectors experience, as well as a pre-test with a mini-inspection. We use sampling to quantify the gain of defects detected from selecting the best inspectors according to the pre-test results compared to the performance of an average team of inspectors. Main findings are that inspector development and quality assurance capability and experience factors do not significantly distinguish inspector groups with different inspection performance. On the other hand the mini-inspection pre-test has considerable correlation to later inspection performance. The sampling of teams shows that selecting inspectors according to the mini-inspection pretest considerably improves average inspection effectiveness by up to one third.

An empirical study on groupware support for software inspection meetings

Software inspection is an effective way to assess product quality and to reduce the number of defects. In a software inspection, the inspection meeting is a key activity to agree on collated defects, to eliminate false positives, and to disseminate knowledge among the team members. However, inspection meetings often require high effort and may lose defects found in earlier inspection steps due to ineffective meeting techniques. Only few tools are available for this task. We have thus been developing a set of groupware tools to lower the effort of inspection meetings and to increase their efficiency. We conducted an experiment in an academic environment with 37 subjects to empirically investigate the effect of groupware tool support for inspection meetings. The main findings of the experiment are that tool support considerably lowered the meeting effort, supported inspectors in identifying false positives, and reduced the number of true defects lost.

Systematically combining process simulation and empirical data in support of decision analysis in software development

Decision-making is a complex and important task in software engineering. The current state-of-the-practice is rather non-systematic as it typically relies upon personal experience without using explicit models. Empirical studies can help but are to some extent context dependent and costly to conduct. Typically it is not efficient or even possible to conduct empirical studies for a large number of context parameter variations. We propose to build on a set of systematic empirical studies to fill gaps in context variable space with simulation: (a) Simulation can use the empirical results from different contexts and apply them to a planning situation as appropriate. (b) The analysis of simulation results can point out situations and factors for which conducting empirical studies would be most worthwhile. This paper presents a general decision model, a simulation framework, and examples for different decisions to use V&V activities in software development (e.g., under which conditions is a V&V activity, such as a re-inspection, worthwhile) to demonstrate practical applications of the general model.

Tailoring a COTS group support system for software requirements inspection

The inspection of early life-cycle artifacts such as requirement documents promises great benefits. However, research demonstrates that the inspection process is complex and expensive and that tool support would be highly desirable. Existing inspection tools focus largely on the inspection of source code. We have therefore devised groupware support for inspecting requirements. Based on our experience with adopting a group support system (GSS) for requirements negotiation, we decided to tailor this commercial GSS to support inspection of requirements. The paper discusses our concept of a Groupware-supported Requirements Inspection Process (GRIP) and shows that tailoring a COTS GSS works well to automate this process.

Teaching the unified process to undergraduate students

A modern software development process framework, such as the unified process, is able to overcome problems of past frameworks, e.g., structural mismatches (notation, method used) between the products in different process steps. Moreover, it promotes an industry standard, which yields instantly tangible benefits from better communication in development teams and further secures investments in development methods based on such a standard. We describe the concept of a one-year undergraduate course in SE based on the unified process and elaborate why, where and how the unified process can be incorporated into the course. The students response to the process was very positive. We argue that the unified process (UP) qualifies very well for being taught to SE students, as it integrates and extends essential SE practices. Another key argument for applying the UP was to integrate the different aspects and topics covered in the SE course and embedding them into a single structured and systematic SE process. By doing so students develop a much better understanding of the relationships and interactions between different models, analysis steps and techniques.