News

Paper accepted at Expert Systems with Application

26.08.2020
We are happy to announce that the article A Multi-Objective Anytime Rule Mining System to Ease Iterative Feedback from Domain Experts was accepted for publication in the Journal Expert Systems with Applications. The article was joint work by Tobias Baum, Steffen Herbold, and Kurt Schneider and discusses a genetic algorithm, that can be used to learn rules in disjunctive normal form, such that experts not only understand the outcome of the learning process, but are directly able to influence the result of the algorithm through the interactive definition of restrictions on the solution space. 

Paper accepted at the ICSME Research Track

10.08.2020
We are happy to announce that our paper Static source code metrics and static analysis warnings for fine-grained just-in-time defect prediction was accepted for publication at the Research Track of the 36th International Conference on Software Maintenance and Evolution.The paper reports the results of a large-scale study of 38 open source projects over 15 years of project history. We investigated whether a fine-grained just-in-time defect prediction approach can benefit from additional static source code features as well as static analysis warnings.It includes an evaluation of different feature sets on two machine learning algorithms, an advanced SZZ algorithm to find bug-inducing changes and a specialized cost model for defect prediction.It is the largest study to date that includes static source code metrics and the first study to include static analysis warning density based features in just-in-time defect prediction.

Article accepted at the 13th System Testing and Validation (STV'20) workshop

10.07.2020
We are happy to announce that our article Using TDL for Standardised Test Purpose Definitions is accepted at the 13th IEEE International Workshop on System Testing and Validation (STV'20), co-located with 20th IEEE International Conference on Software Quality, Reliability, and Security (QRS 2020). Within this article, we discuss the Structured Test Objective Specification extension for the Test Description Language (TDL) and its application in several contexts related to standardised test purposes.

Philip Makedonski elected as member of the board of the Institute of Computer Science

10.07.2020
We are happy to announce that Dr. Philip Makedonski has been elected as a member of the board of the Institute of Computer Science.

Registered Report accepted at the MSR

01.04.2020
We are happy to announce that our study protocol Large-Scale Manual Validation of Bugfixing Changes has been accepted at the Registered Reports track of the 2020 IEEE/ACM 17th International Conference on Mining Software Repositories (MSR 2020). The protocol describes how we want to validate a large amount of bug fixing commits to create a new and detailed data set about bug fixing changes that can be used by researchers for different purposes, e.g., defect prediction, program repair, or bug localization. Our goal is to conduct this project together with the community and we invite other researchers to contribute. The protocol is already registered in the OSF Registry

Paper accepted at the ICSE Demo Track

03.02.2020
We are happy to announce that our paper The SmartSHARK Ecosystem for Software Repository Mining was accepted for publication at the Demostrations Track of the 42nd International Conference on Software Engineering. The paper presents our SmartSHARK ecosystems that is comprised of tools for various software mining tasks, e.g., the scraping of data from the Git version control system, the Jira issue tracker, or mailing lists. SmartSHARK can also collect data the source code, e.g., software metrics, detected refactorings, or the types of changes. Additional tools in the ecosystem allow the application of heuristics and manual validation to further enrich the collected data. SmartSHARK exposes all data for analysis and visualization in a single database, that can be used by researchers to conduct experiments. The International Conference on Software Engineering is the most important conference in the field of software engineering and, together with the top journals in the field, one of the most reputable venues for research publications. 

Philip Makedonski elected as Vice-Chair of ETSI TC MTS

29.01.2020
We are happy to announce that Dr. Philip Makedonski has been elected as vice-chair of the Technical Committee Methods for Testing and Specification (TC MTS) at the European Telecommunications Standards Institute (ETSI) for the period of two years until January 2022.

Paper accepted at the ICSE New Ideas Track

16.01.2020
We are happy to announce that Steffen Herbold's paper With Registered Reports Towards Large Scale Data Curation was accepted for publication at the New Ideas and Emerging Results track of the 42nd International Conference on Software Engineering. The paper proposes an approach for crowd sourcing research tasks based on the concept behind Amazon's mechanical turk. The critical differences are that the research projects are in the form of pre-registered studies and that the motivation for the contribution to the crowd working is not monetary, but credit in the form of authorship. The International Conference on Software Engineering is the most important conference in the field of software engineering and, together with the top journals in the field, one of the most reputable venues for research publications. 

Happy Holidays! And best wishes for 2020

20.12.2019
The software engineering for distributed systems group wishes everyone a joyous holiday season, Merry Christmas and a happy and peaceful New Year.

Paper accepted at the IEEE Transactions on Software Engineering

02.12.2019
We are happy to announce the Steffen Herbold's article On the cost and profit of software defect prediction was accepted for publication by the IEEE Transactions on Software engineering. The article defines a cost model for software defect prediction. Based on the cost model, we derive required upper and lower boundaries that must be fulfilled in order for a defect prediction model to have any chance of saving costs. As part of the analysis of costs, we determined a systematic problem in nearly the complete body of literature of defect prediction: we ignore the n-to-m relationship between artifacts and defects and usually assume a 1-to-1 or 1-to-m relationship (binary labels, defect counts). However, costs of defect prediction models can vary a lot if the n-to-m relationship accounted for. 

Pages

2024 © Software Engineering For Distributed Systems Group

Main menu 2