Predicting Bugs Using Antipatterns - PowerPoint Presentation

Slide1

Predicting Bugs Using Antipatterns

Ehsan Salamati Taba, Foutse Khomh, Ying Zou, Meiyappan Nagappan, Ahmed E. Hassan

1

Slide2

2

Slide3

Predict BugsModel

Code

Antipatterns

3

Past Defects, History of Churn (Zimmermann, Hassan et al.)

Topic Modeling (Chen et al.)

Slide4

4

Slide5

not technically incorrect and don't prevent a system from functioning

weaknesses in design5

Antipatterns

Slide6

Indicate a deeper problem in the system

6

Slide7

Antipatterns indicate weaknesses in the design that may increase the risk for bugs in the future. (Fowler 1999)

Motivation7

Slide8

There is not a lot of refactoring activities when developing a system. (Olbrich et al.)

Motivation8

Slide9

Approach

CVS Repository

Mining Source Code Repositories

Detecting

Antipatterns

Mining Bug Repositories

Bugzilla

Calculating Metrics

Analyzing

RQ1

RQ2

RQ3

9

Slide10

10SystemsRelease(#)Churn

LOCsEclipse2.0 - 3.3.1(12)148,45426,209,669ArgoUML

0.12 - 0.26.2(9

)

21,427

2,025,730

Studied Systems

Studied Systems

Mining Source Code Repositories

Slide11

Approach

CVS Repository

Mining Source Code Repositories

Detecting

Antipatterns

Mining Bug Repositories

Bugzilla

Calculating Metrics

Analyzing

RQ1

RQ2

RQ3

11

Slide12

Detecting Antipatterns

1213 different antipatternsDECOR (Moha et al.)# of Antipatterns

# Files

Systems

#

Antipatterns

Eclipse

273,766

ArgoUML

15,100

Slide13

Approach

CVS Repository

Mining Source Code Repositories

Detecting

Antipatterns

Mining Bug Repositories

Bugzilla

Calculating Metrics

Analyzing

RQ1

RQ2

RQ3

13

Systems

#Post Bugs

#Pre Bugs

Eclipse

27,406

23,554

ArgoUML

2,549

2,569

Slide14

Research QuestionsRQ1: Do antipatterns affect the density of bugs in files?RQ2:

Do the proposed antipattern based metrics provide additional explanatory power over traditional metrics?RQ3: Can we improve traditional bug prediction models with antipatterns information?14

Slide15

RQ1: Do antipatterns affect the density of bugs in files?

Null HypothesisDensity of bugs in the files with antipatterns and the other files without antipatterns is the same.15

Wilcoxon rank sum test

Slide16

16SystemsReleases(#)D

A – DNA> 0p-value<0.05Eclipse1288ArgoUML966

Files with

Antipatterns

Density of Bugs

Files without

Antipatterns

Density of Bugs

RQ1:

Do

antipatterns

affect the density of bugs in files?

Slide17

Research QuestionsRQ1: Do antipatterns

affect the density of bugs in files?RQ2: Do the proposed antipattern based metrics provide additional explanatory power over traditional metrics?RQ3: Can we improve traditional bug prediction models with antipatterns information?17

Slide18

RQ2: Metrics Average Number of Antipatterns

(ANA) Antipattern Cumulative Pairwise Differences (ACPD)18

Antipattern Recurrence Length(ARL)

Antipattern

Complexity Metric (ACM)

Slide19

19

1.0

2

.0

3

.0

4

.0

5

.0

6

.0

a.java

b

.java

c

.java

3

4

0

2

1

3

4

5

1

0

0

3

0

6

5454ANA(a.java) =2.16, ARL(a.java) = 18.76, ACPD(a.java) = 0RQ2: Example

Slide20

20

Slide21

21

Provide

additional explanatory power over traditional

metrics

ARL shows the biggest improvement

Slide22

Research QuestionsRQ1: Do antipatterns affect the density of bugs in files?

RQ2: Do the proposed antipattern based metrics provide additional explanatory power over traditional metrics?RQ3: Can we improve traditional bug prediction models with antipatterns information?22

Slide23

RQ3: Can we improve traditional bug prediction models with

antipatterns information?Intra System ModelsStep-wise analysisRemoving Independent VariablesCollinearity Analysis

23

Metric name

Description

LOC

Source lines of codes

MLOC

Executable lines of codes

PAR

Number of parameters

NOF

Number of attributes

NOM

Number of methods

NOC

Number of children

VG

Cyclomatic

complexity

DIT

Depth of inheritance tree

LCOM

Lack of cohesion of methods

NOT

Number of classes

WMC

Number of weighted methods per class

PRE

Number of pre-released bugs

Churn

Number of lines of code addedmodified or deleted

Slide24

24

ARL remained statistically significant and had a low

collinearity

with other metrics

# Versions

# Versions

Slide25

RQ3: Can we improve traditional bug prediction models with

antipatterns information?F-measure25

ARL can improve cross-system bug prediction on the two studied systems

Slide26

Slide27

Backup Slides

27

Slide28

28

1.0

2

.0

3

.0

4

.0

5

.0

6

.0

a.java

b

.java

c

.java

3

4

0

2

1

3

4

5

1

0

0

3

0

6

5454ANA(a.java) =2.16, ARL(a.java) = 18.76, ACPD(a.java) = 0RQ2) Example

Slide29

29Anti Singleton

BlobClass Data Should be Private

Complex

Class

Large Class

Lazy Class

LPL

Long Method

Message Chain

RPB

Spaghetti

Code

SG

SwissArmy

Knife

-

-

Slide30

30

Slide31

RQ1) Do

antipatterns affect the density of bugs in files?HypothesisThere is no difference between the density of future bugs of the files with antipatterns and the other files without antipatterns.Wilcoxon rank sum testWe perform a Wilcoxon rank sum test to acceptor refuse the hypothesis, using the 5% level (i.e., p-value < 0:05).

Hypothesis

There is no difference between the density of future bugs of the files with

antipatterns

and the other files without

antipatterns

.

Wilcoxon rank sum

test

Findings

In general, the density of bugs in a file with

antipatterns

is higher than the density of bugs in

a file

without

antipatterns

.

31