Bug(fault)를 예측하려는 노력

Post on 16-Jul-2015

66 views 2 download

Transcript of Bug(fault)를 예측하려는 노력

Fault(bug)를 예측하려는 노력

(Quantitative Analysis of Fault Distributions in Complex Software Systems)

SEEG

김진태

Fault, bug, defect, error, failure를 구별하실 수 있으세요?

Fault를 예측하면 무엇이 좋을까요?

Fault를 예측할 수 있을까요?

오늘 발표는 fault를 예측해보고자 하는 눈물겨운이야기 입니다.

2000년 8월, TSE

Norman Fenton

Professor, Queen Maryand Westfield College, London

Niclas Ohlsson

Ph.D, Technical director, GratisTelInternational

2007년 5월, TSE

Carina Andersson

research associate in the Department ofComputer Science, Lund University, Sweden

Per Runeson

Professor, softwareengineering at Lund University, Sweden

2013년 4월, TSE

Tihana GalinacGrbac

Professor, University of Rijeka, Croatia

Per Runeson

Professor, softwareengineering at Lund University, Sweden

Darko Huljenic

Professor, University of Zagreb, Croatia

실험을 대상 프로젝트들의 현황

실험을 대상 Module의 현황

Hypothesis 1a. A small number of modules contain most of the faults detected during prerelease testing.

Pre-release post-release

YESHypothesis 2a. A small number of modules contain most of the faults detected during postrelease testing.

YES

Hypothesis 1b. If a small number of modules contain most of the prerelease faults, then it is because these modules constitute most of the code size.

NOHypothesis 2b. If a small number of modules contain most of the postrelease faults, then it is because these modules constitute most of the code size.

NO

Hypothesis 3. Higher incidence of faults in FT implies higher incidence of faults in ST.

YES

Hypothesis 4. A higher incidence of faults in prerelease testing implies higher incidence of faults in postrelease.

YES

Hypothesis 5a. Smaller modules are less likely to be failureprone than larger ones.

Hypothesis 5b. Size metrics are good predictors of prereleasefaults in a module.

Hypothesis 5c. Size metrics are good predictors of postreleasefaults in a module.

Hypothesis 5d. Size metrics are good predictors of a module’sprerelease fault density.

NO

NO

NO

NO

정리