An Assesment of Visualization of Representative Class Interaction Scenarios for Program Comprehension

Transcription

1 An Assesment of Visualization of Representative Class Interaction Scenarios for Program Comprehension Satoshi Munakata, 1 Yui Watanabe, 1 Takashi Ishio 1 and Katsuro Inoue 1 In object-oriented programming, it is effective to visualize an execution trace of an instance of a class as behavior diagrams for understanding behaviors of the class. However, a class can create a large number of instances; it is hard for developers to read behavior diagrams for each instance. In this paper, we propose a method to classify instances into groups based on their behaviors, and visualize only representative behaviors of their instances. Users can select one of four categorization criteria depends on their purpose of investigation for program comprehension. As an assessment of the efficacy for program comprehension, we have applied our approach to six Java programs. 1. 9) 1 2),3),5),9),11) 7),8) UML 12) UML DOPG UML 4 GUI Graduate School of Information Science and Technology, Osaka University 1 c 2010 Information Processing Society of Japan

2 ) 2),3),5),9),11) Quante 1 DOPG(Dynamic Object Process Graph) 5) DOPG DOPG 8 (Read Write Call Entry Return Start Final Condition) 3 (Conditional Unconditional Invocation) Lange 4) ),8) Salah 8) Richner 7) ID ID ( ) Use(o) : o ( 1(a)) Used(o) : o ( 1(b)) Methods(o) : o ( 1(c)) Called(o) : o ( 1(d)) c 2010 Information Processing Society of Japan

3 1 Use Used Methods Called DOPG UML E use (o s,o k ) 2 o s,o k E use(o s,o k ) E use (o s,o k ) Use(o s )=Use(o k ) 1),6) 2 E use (o s,o k ) E used (o s,o k ) 2 o s,o k E used (o s,o k ) E used (o s,o k ) Used(o s )=Used(o k ) 2 E used (o s,o k ) E methods (o s,o k ) 2 o s,o k E methods (o s,o k ) E methods (o s,o k ) Methods(o s)=methods(o k ) 2 E methods (o s,o k ) E called (o s,o k ) 2 o s,o k E called (o s,o k ) E called (o s,o k ) Called(o s)=called(o k ) 2 E called (o s,o k ) E use (o s,o k ) O G use (O) ={G 1,,G l } G i G use (O) E methods G use,methods (G i)={g i1,,g im} E use,e methods DOPG UML 1 4. Amida-OGAN 2 GUI Explorer Metrics View Feature View Visualize View GUI Java c 2010 Information Processing Society of Japan

4 2 Software #Events #Classes #Objects Scheduler 8578 (12) JHotDraw 128,796 (97)185 6,286 Apache PMD 633,309 (67)105 8,431 MASU 3,198,125 (80)309 12,414 Apache FOP 3,657,813 (120)502 27,342 Antlr 7,305,687 (42)60 21,022 14,932,308 (418)1,178 77,468 2 Amida-OGAN ( 2 ) java javax sun com.sun %( 36%) (1) (2) (3) (1)(2) (3) 5.1 :Scheduler Scheduler ( ) GUI CalendarDate Scheduler java javax sun com.sun CalendarDate E use 3 {s 1,s 2,s 3 } s s 1 87 s s 1 CalendarMonth s 2 s 3 3 s 3 4 RegisterDialog CalendarDate s 1,s 2,s 3 DOPG. 1 DOPG 4 c 2010 Information Processing Society of Japan

5 3 CalendarDate {s 1,s 2,s 3} s 2 s 3 4 s3 CalendarDate CalendarMonth CalendarDate 18 s 2 s 3 s 2 DOPG DateCellRenderer DateCell CalendarDate s 2 UML s 2 s 3 DOPG s 2 3 s 2,s 3 s 3 DOPG RegisterDialog UML s 3 CalendarDate CalendarDate CalendarDate CalendarDate CalendarDate s 1 s 2 s 3 E use CalendarDate GUI 1 CalendarDate 731 CalendarDate 3 s 1,s 2,s CalendarDate CalendarDate DOPG DOPG 6 (1) (2)2 5 c 2010 Information Processing Society of Japan

6 :MASU MASU MASU Java C++ C# UnresolvedMethodIndo AST 84 E methods 4 AST 4 DOPG DOPG DOPG DOPG %( 2.4%) ( 1.6 ) 3 DOPG DOPG 4 3 Software #E use #E Used #E methods #E called # #DOPG Scheduler JHotDraw ,657 Apache PMD MASU ,188 Apache FOP ,928 Antlr ,091 2,291 1,634 1,785 1,658 2,525 11,421 5 E use 6 E used ,6,7,8 E use,e used,e methods,e called 5,6,7,8 5,6,7,8 9 E use 9 97% 50 82% E used 99% 95% E methods 99% 87% E called 99% 94% 6 c 2010 Information Processing Society of Japan

7 7 E methods 8 E called 9 E use E methods 10 E methods (a) (b) (a) MASU StateChangeEvent NamespaceInfo JHotDraw ResourceBundleUtil AttributeAction E use MASU StateChangeEvent 8344 E use E use E methods 9 E use E methods 9 E use E methods StateChangeEvent 8344 E methods 3 DOPG UML E use 9 7 c 2010 Information Processing Society of Japan

8 10 E methods ( ) 10 (b) 10 ( 1 ) (1) (2) Scheduler 3 3 (3) % 9 1 1) G.Arevalo, F.Buchli, and O.Nierstrasz. Detecting implicit collaboration patterns. In the 11th Working Conference on Reverse Engineering, pp , ) V.Dallmeier, C.Lindig, A.Wasylkowski, and A.Zeller. Mining object behavior with adabu. In the 4th International Workshop on Dynamic Analysis, pp , ) T. Gschwind and J. Oberleitner. Improving dynamic data analysis with aspectoriented programming. In the 7th European Conference on Software Maintenance and Reengineering, pp , ) D.B. Lange and Y.Nakamura. Interactive visualization of design patterns can help in framework understanding. In ACM SIGPLAN Notices, Vol.30, pp , ) J.Quante and R.Koschke. Dynamic object process graphs. In Journal of Systems and Software, Vol.81, pp , ) T.Richner and S.Ducasse. Recovering high-level views of object-oriented applications from static and dynamic information. In the 15th International Conference on Software Maintenance, pp , ) T. Richner and S. Ducasse. Using dynamic information for the iterative recovery of collaborations and roles. In the 18th International Conference on Software Maintenance, pp , ) M.Salah, T.Denton, S.Mancoridis, A.Shokoufandeh, and F.I. Vokolos. Scenariographer: A tool for reverse engineering class usage scenarios from method invocation sequences. In the 21st International Conference on Software Maintenance, pp , ) T.Systa. Understanding the behavior of java programs. In the 7th Working Conference on Reverse Engineering, pp , ) N. Wilde and R. Huitt. Maintenance support for object-oriented programs. In IEEE Transactions on Software Engineering, Vol.18, pp , ) T. Xie and D. Notkin. Automatic extraction of sliced object state machines for component interfaces. In the 3rd Workshop on Specification and Verification of Component-Based Systems, pp , ),,.. 163, 31, pp , : c 2010 Information Processing Society of Japan