Software Quality Talks

Thursday, March 17, 2011
13.30-14.30, BIN 1.D.22 (new location!) former: BIN 0.B.06

Abstract
Due to their pivotal role in software engineering, software engineers spend considerable effort on the quality assurance of software requirements specifications. As they are mainly described in natural language, relatively few means of automated quality assessment exist. We found, however, that clone detection, a technique widely applied to source code, is promising to assess one important quality aspect in an automated way, namely redundancy that stems from copy&paste operations. I describe in this talk a large-scale case study that applied clone detection to 28 requirements specifications with a total of 8,667 pages. I report on the amount of redundancy found in real-world specifications, discuss its nature as well as its consequences, and evaluate in how far existing code clone detection approaches can be applied to assess the quality of requirements specifications in practice.

Short Bio
Stefan Wagner received a diploma in computer science from the University of Applied Sciences, Augsburg, an MSc in distributed and multimedia information systems from Heriot-Watt University, Edinburgh, and a doctoral degree from Technische Universität München. Since 2007, he has been working at TU München in the Software Systems Engineering group of Prof. Broy heading the competence centre fro software quality. He leads the consortiums project Quamoco in which he develops jointly with partners like SAP and Siemens a new quality model. Dr. Wagner has published more than 50 national and international contributions to quality specification and evaluation, software testing, clone detection, and requirements engineering.

Thursday, March 17, 2011
16.00-17.00, BIN 2.A.10

Abstract
Reliably predicting software defects is one of the holy grails of software engineering. Researchers have devised and implemented a plethora of defect prediction approaches, which vary in terms of accuracy, complexity, and the input data they require. However, the absence of an established benchmark makes it hard, if not impossible, to compare approaches. We discuss a benchmark for defect prediction, which we exploit to compare well-known techniques, together with novel approaches we devised. We measure the predicting performances in two scenarios: entity classification (defect-prone or not) and ranking. We also take into account the effort needed to review the entities. Based on the results of our comparison, we illustrate a number of insights on the prediction approaches, and we outline future research directions in the field.

Short Bio
Marco D'Ambros is a postdoctoral researcher at the University of Lugano, Switzerland. He earned his PhD in software engineering from the same University in October 2010, under the supervision of Prof. Michele Lanza. He received MSc degrees from both Politecnico di Milano (Italy) and the University of Illinois at Chicago. His research interests lie in the domain of software engineering with a special focus on software evolution, software visualization, and defect prediction & analysis. He authored more than 25 technical papers, and is the creator of several software visualization and program comprehension tools.

Friday, March 18, 2011
08.30-09.30, BIN 1.B.18

Abstract
In the development of a software system, large amounts of new information are produced continuously. Source code, bugs, iteration plans and documentation, to name just a few, are changed or newly created by developers of the software system every day. As a developer works on the system, she does not only produce or change the information in such artifacts, she also has to find information to answer questions and stay aware of the relevant information. However, to effectively complete her task, the developer only requires the small portion of information that is absolutely pertinent to her work. To support developers in coping with this overload and managing the information, we propose two developer-centric models. The information fragment model supports the automatic integration of information to help rank, filter and interpret the information a developer might be interested in. The Degree-of-Knowledge (DOK) model provides a means to automatically determine the core of what a developer knows. This knowledge model can then be used to identify the information a developer might be interested in.

Short Bio
Thomas Fritz is a Ph.D. candidate in the Department of Computer Science at the University of British Columbia. He completed his Diplom thesis as part of the OBASCO (Objects, Aspects and Components) group at the Ecole des Mines de Nantes, France, and completed his Diplom degree at the Ludwig-Maximilians-University Munich, Germany, in 2005. He also has experience working as an intern with several companies including the IBM labs in Zurich and Ottawa. His research focuses on how to help software developers better manage the information and systems on which they work. At ICSE 2010, Thomas won an ACM SIGSOFT Distinguished Paper Award and came third in the ACM Student Research Competition.

Friday, March 18, 2011
11.00-12.00, BIN 1.B.18

Abstract
In the last decade, the use of open source software has become an important factor to reduce the costs of IT projects. Unfortunately, the success of many projects is hindered by the lack of documentation. Many libraries are difficult to use because implicit rules (such as a call to close() must be preceded by a call to open()) are undocumented. To approach this problem, specification mining aims at learning such specifications from program executions. In this talk, I will take two steps towards making mined specifications applicable in practice. First, I will introduce object behavior models, a specification mining technique that yields concise specifications for the behavior of individual objects. Second, I will approach the problem of incomplete specifications by combining test case generation with specification mining.

Short Bio
Valentin Dallmeier studied computer science at the universities of Passau and Saarbrücken (Germany). He received his diploma in 2005 from Saarland University and joined the chair for software engineering lead by Andreas Zeller in the same year. His research interests are applications of dynamic program analysis, in particular fault localization and mining specifications. In 2010, his dissertation work was awarded the Ernst Denert award for software engineering.

Friday, March 18, 2011
14.30-15.30, BIN 1.B.18

Abstract
In this talk I argue that we need a radical change in the way we approach software assessment both in practice and in research. Assessment is a critical software engineering activity, often accounting to as much as 50% of the overall development effort. However, in practice this activity is regarded as secondary and it is dealt with in an ad-hoc way. This does not service. We should recognize it explicitly and approach it holistically as a discipline. Why holistically? Because software evolution is a multidimensional phenomenon that exhibits itself in multiple forms and at multiple levels of abstraction. For example, software evolution spans over multiple topics such as modeling, data mining, visualization, human-computer interaction, or even language design. There exists an extensive body of research in each of these areas, but these approaches are mostly disparate and thus have little overall impact. We need a new kind of research effort that deals with their integration. Ultimately, assessment is a human activity that concerns taking decisions in specific situations. Thus, to be effective, assessment must go beyond general technicalities and deal with those specific situations. For example, instead of having a predefined generic tool, we should be able to craft one that deals with the constraints of the system under study. To accommodate the scale of the problem, the research methods should be adapted to the task as well. First, it is critical to integrate tool building into the research process, because without scalable tools we cannot handle large. Second, we have to work collaboratively both to integrate our conceptual approaches and to share the practical costs of tool building.

Short Bio
Tudor Gîrba attained his PhD in 2005 from the University of Berne, Switzerland, and he now works as an independent consultant. His main expertise lies in the area of software engineering with focus on software and data assessment. Among others, since 2003 he leads the work on the Moose analysis platform (http://moosetechnology.org). He published all sorts of peer reviewed publications, he served in program committees for several dozen international venues, and he is regularly invited to give talks and lectures. He is currently advocating that assessment must be recognized as a critical software engineering activity. He coined the term "humane assessment" (http://humane-assessment.com), and he is currently helping companies to assess and to manage large software systems and data sets.