The paper "FAST Approaches to Scalable Similarity-based Test Case
Prioritization" by Breno Miranda (UFPE, Brazil), Emilio Cruciani (GSSI, Italy), Roberto Verdecchia
(GSSI, Italy, and Vrije Universiteit Amsterdam, NL) and Antonia Bertolino (ISTI - CNR, Italy) has been accepted as a technical research paper at ICSE 2018, the 40th International Conference on Software Engineering. ICSE is the flagship conference in Software Engineering and is very selective its technical-research-paper track is the most prestigious one within the conference.
The paper contributes to the classic area of software testing, which is one of the approaches developed by computer scientists to increase their confidence that computing systems actually do what they were designed to achieve. Since the number of tests that can be performed on a computing system is enormous, test-case prioritization is a crucial element in any practical testing framework. In that approach, one prioritizes test cases so that they can detect faults more efficiently using the available limited resources.
The paper to be presented at ICSE 2018 is the first that applies techniques from data mining to test case prioritization. In particular, it shows that the use of ideas from locality-sensitive hashing, a technique stemming from research in TCS that has been employed to great effect in approximate similarity searches in audio and video data, amongst others, leads to effective test prioritization in practice, when one needs to select tests effectively amongst millions of possible ones.
Antonia Bertolino is a member of the Scientific Board for the PhD programme in Computer Science at the GSSI. Breno Miranda is currently a postdoctoral researcher in Brazil and is one of Antonia Bertolino's former PhD students. Emilio Cruciani and Roberto Verdecchia just started their second year as PhD students in computer science at the GSSI and the paper to be presented at ICSE 2018 builds on their project for the first-year Software Testing course held in early 2017 at the GSSI by Antonia Bertolino. The project itself arose from a question asked by the students during the lectures. This is what inspiring, research-based teaching can produce when there is intellectual chemistry between lecturers and students.
The paper contributes to the classic area of software testing, which is one of the approaches developed by computer scientists to increase their confidence that computing systems actually do what they were designed to achieve. Since the number of tests that can be performed on a computing system is enormous, test-case prioritization is a crucial element in any practical testing framework. In that approach, one prioritizes test cases so that they can detect faults more efficiently using the available limited resources.
The paper to be presented at ICSE 2018 is the first that applies techniques from data mining to test case prioritization. In particular, it shows that the use of ideas from locality-sensitive hashing, a technique stemming from research in TCS that has been employed to great effect in approximate similarity searches in audio and video data, amongst others, leads to effective test prioritization in practice, when one needs to select tests effectively amongst millions of possible ones.
Antonia Bertolino is a member of the Scientific Board for the PhD programme in Computer Science at the GSSI. Breno Miranda is currently a postdoctoral researcher in Brazil and is one of Antonia Bertolino's former PhD students. Emilio Cruciani and Roberto Verdecchia just started their second year as PhD students in computer science at the GSSI and the paper to be presented at ICSE 2018 builds on their project for the first-year Software Testing course held in early 2017 at the GSSI by Antonia Bertolino. The project itself arose from a question asked by the students during the lectures. This is what inspiring, research-based teaching can produce when there is intellectual chemistry between lecturers and students.
Congratulations to the authors (and to the computer science group at the GSSI)!
No comments:
Post a Comment