Probabilistic Fault Localisation

David Landsberg*, Hana Chockler, Daniel Kroening

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingChapterpeer-review

2 Citations (Scopus)
201 Downloads (Pure)

Abstract

Efficient fault localisation is becoming increasingly impor- tant as software grows in size and complexity. In this paper we present a new formal framework, denoted probabilistic fault localisation (pfl), and compare it to the established framework of spectrum based fault localisation (sbfl). We formally prove that pfl satisfies some desirable properties which sbfl does not, empirically demonstrate that pfl is sig- nificantly more effective at finding faults than all known sbfl measures in large scale experimentation, and show pfl has comparable efficiency. Results show that the user investigates 37% more code (and finds a fault immediately in 27% fewer cases) when using the best performing sbfl measures, compared to the pfl framework. Furthermore, we show that it is theoretically impossible to design strictly rational sbfl measures that outperform pfl techniques on a large set of benchmarks.
Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Subtitle of host publicationHardware and Software: Verification and Testing
PublisherSpringer‐Verlag Berlin Heidelberg
Pages65-81
Number of pages17
Volume10028 LNCS
ISBN (Print)9783319490519
DOIs
Publication statusPublished - 1 Nov 2016
Event12th International Haifa Verification Conference, HVC 2016 - Haifa, Israel
Duration: 14 Nov 201617 Nov 2016

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10028 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Conference

Conference12th International Haifa Verification Conference, HVC 2016
Country/TerritoryIsrael
CityHaifa
Period14/11/201617/11/2016

Keywords

  • Fault localisation
  • Spectrum based fault localisation
  • Triage and debug technologies

Fingerprint

Dive into the research topics of 'Probabilistic Fault Localisation'. Together they form a unique fingerprint.

Cite this