1. Quaid-i-Azam University
Refactoring Patterns Study in Code Clones during Software Evolution
Jaweria Kanwal
Quaid-i-Azam University
Islamabad, Pakistan
Katsuro Inoue
Osaka University
Osaka, Japan
Onaiza Maqbool
Quaid-i-Azam University
Islamabad, Pakistan

2. Introduction Code Clones
Similar code fragments in software
Types of clones
Type1, Type2, Type3
Sometimes clones cause an additional maintenance effort
change in one clone fragment may cause change in other clone fragments [1]
One of the maintenance tasks is to remove the clones from the system through refactoring [2]
Clones need to be detected for better software maintenance 2 2

3. Clones refactoring Refactoring Fowler’s refactoring patterns [3]
Widely used technique to improve code quality
Fowler’s refactoring patterns [3]
Some examples
Move method
Extract method
Extract interface
Add parameter
Clone refactoring may/may not remove them from the system 3

4. Motivation Existing work
Study of changes behaviour of clones in versions [4]
Consistent and inconsistent change
Disappear in few check-ins
All clones are not refactorable
There is no work on the study of actual refactoring tasks performed on code clones 4

5. Research questions How often code clones refactored in software?
How often clones refactored consistently? 5

6. Study Approach 6 JHotDraw Guava Jabref CloneMiner JFreeChart
Xerces_J
CloneMiner
a clone detection tool
RefFinder
Detection of refactorings from consecutive versions
An application is developed in C# for mapping 6

7. How often code clones refactored in software?7

8. Average clone refactorings in five Java systems
Versions
JHotDraw
Guava
Jabref
JFreeChart
Xerces_J Vn
27. 4%
47.0%
6.0%
3.8%
32.0%
Vn+1
19.1%
0.8%
9.0%
4.5%
24.8%
Vn+2
21.1%
0.5%
4.8%
1.3%
34.0%
Vn+3
21.4%
17.0%
34.2%
Vn+4
33.3%
6.2%
6.7%
1.2%
2.0%
Vn+5
8.2%
0.4%
0.6%
10.5%
Vn+6
3.4%
7.1%
Vn+7
Average
24.2%
Release time between version Vn+2 and Vn+3 is four months and release duration between version Vn+3 and Vn+4 is only one month
Actual refactorings are 496 in Vn+3 and only seven in next version 8

9. How often clones refactored consistently?9

10. Refactoring consistency in Versions
Clone class
Add_parameter
Add_parameter I1 I1 I1 I1
Add_parameter I2 I2 I2 I2
Add_parameter I3 I3 I3
Add_parameter I3 I4 I4 Vn
Vn+1 Vn
Vn+1
(a) An example of consistent refactoring
(b) An example of in- consistent refactoring 10

11. Percentage of consistent refactoring in five Java systems
Versions
JHotDraw
Guava
Jabref
JFreeChart
Xerces_J Vn
62.7%
62.0%
50.0%
32.0%
53.1%
Vn+1
56.7%
20.0%
80.0%
42.0%
40.0%
Vn+2
73.5%
100%
42.8%
47.1%
Vn+3
52.0%
45.4%
70.4%
42.1%
49.0%
Vn+4
66.8%
45.0%
32.3%
44.1%
Vn+5
48.0%
10.0%
36.0%
34.2%
Vn+6
45.1%
44.4%
Vn+7
52.3%
Average
62.2%
53.2%
47.2%
38.0% 11

12. Conclusions and Future work
A small portion of code clones are refactored
Clone refactoring depends upon actual refactoring tasks performed during the releases
Consistent refactorings
More than 40% clones are refactored consistently in most of the versions
Future work
Consequences of inconsistent clone refactorings on software maintenance 12

13. References
M. Kim, V. Sazawal, D. Notkin, and G. Murphy. An Empirical Study of Code Clone Genealogies. ACM SIGSOFT Software Engineering Notes, 30(5), pp , 2005.
M. Fowler, Refactoring: Improving the Design of Existing Programs. Addison-Wesley, 1999.
H. Basit, and S. Jarzabek. A data mining approach for detecting higher-level clones in software. In journal of IEEE Transactions on Software Engineering, 35(4), pp , 2009.
Kamiya, Toshihiro and Kusumoto, Shinji and Inoue, Katsuro. CCFinder: a multilinguistic token-based code clone detection system for large scale source code, IEEE Transactions on Software Engineering, volume={28}, number={7}, Pp , 2002, IEEE.
Aversano, Lerina and Cerulo, Luigi and Di Penta, Massimiliano. How clones are maintained: An empirical study,11th European Conference on Software Maintenance and Reengineering, pp , 2007, IEEE.
Mens, Tom and Tourwe’, Tom. A survey of software refactoring, IEEE Transactions on software engineering, volume={30}, number={2}, Pp , 2004, IEEE. 13

14. ?