Towards Well Planned Code Cities First Name: Natalia Last Name: Tymchuk Type: Talk Abstract: Software visualization has been shown as effective to support program and system comprehension. A possible solution adopts the city metaphor, and visualizes software systems as 3D cities. Most of the existing implementations, like CodeCity, exploit rectangle packing layout for placing the entities in a compact positioning. However, the distance between elements is determined by the layout algorithm only, without any connection to the actual structure of the system and the relations between software entities, which are fundamental for comprehension and maintenance. We are currently investigating a new 3D representation for visualizing large software systems as cities. In our tool, we exploit force based layout to place software entities like classes and packages based on their relations, for example by considering client dependency. Our cities look more ``organic'', characterized by dependent entities grouped together, and less coupled entities more far away from each other. Bio: I'm a Ph.D. student at the Faculty of Informatics of University of Lugano. I am working under the supervision of Prof. Dr. Michele Lanza, in the REVEAL research group. My main topic is how to improve the visualization of software systems in a useful and realistic way. On the last ESUG I presented Voronoi diagram builder, while 2 years ago I worked on the project “SciSmalltalk - Ordinary Differential Equations Solver” in GSoC.

1. Natalia Tymchuk, Michele Lanza
Towards Well Planed
Code City
R AE E LV

2. CodeCity

3. Class
Package or Namespace
Number of
attributes
CodeCity
Number of
methods

5. Manhattan

6. Barcelona

7. Paris

8. Lviv

9. Voronyj Diagram

14. 1 minute
2 minutes
3 minutes
4 minutes
Number of Sites
0 50 100 150 200 250

15. 1 minute
2 minutes
3 minutes
4 minutes
Number of Sites
0 250 500 1000 1500 2000 2500

20. (x,y)?
(x,y)?
(x,y)?(x,y)?
(x,y)?
(x,y)?

21. 1 2 3 4 5 6 7 8
1 x
2 x x
3 x x
4 x
5 x x
6 x
7 x
8 x

22. 1 2 3 4 5 6 7 8
1 x
2 x x
3 x x
4 x
5 x x
6 x
7 x
8 x

23. 1 2 3 4 5 6 7 8
1 x
2 x x
3 x x
4 x
5 0 x 0 0 0 x 0 0
6 x
7 x
8 x

24. 1 2 3 4 5 6 7 8
1 x
2 x x
3 x x
4 x
5 0 x 0 0 0 x 0 0
6 x
7 x
8 x
(a,b)

25. Multidimensional scaling

26. Multidimensional scaling

29. Force based layout

30. Multidimensional scaling Force based layout

31. 1 method 60 methods

32. 38 methods1 method

41. Graph model;
Different features for software entities;
Distance, position;

43. New view on Force based layout

44. Nodes’ Force

45. Nodes’ Force

46. Springs’ Force

47. Springs’ Force

48. Resulting Force

49. RepulsionValue
Mass
Repulsion Radius
Stiffness
Ideal Length

50. Node: RepulsionValue

51. Node: RepulsionValue

52. 10 100 1000
50
100
500
1000
Repulsion Value Iterations
Node: RepulsionValue

53. Node: Mass

54. Node: Mass

55. 10 100 1000
5
20
40
Mass Iterations
Node: Mass

56. Node: Repulsion Radius

57. Node: Repulsion Radius

58. Node: Repulsion Radius

59. Node: Repulsion Radius

60. Node: Repulsion Radius

61. 10 100 1000Iterations
Node: Repulsion Radius

62. Edge: Stiffness

63. Edge: Stiffness

64. Stiffness 0.1
initial length 200
Stiffness 10
ideal length 10
repulsion 50
iterations 50

65. Edge: Ideal length

66. Edge: Ideal length
10
20

67. Ideal length 0
initial length 200
Ideal length 10
stiffness 0.1
repulsion 1
iterations 50

68. Features:
Distance
Time

69. Distance

70. Distance

71. Distance

72. 100 500 1000
Iterations
1
Traditional
distance
Padded
distance
Distance

73. Slowing down the time

74. Slowing down the time

75. Slowing down the time

76. Slowing down the time

77. Slowing down the time

79. RTForceBasedLayout RTFBLayout

80. RTFBLayout
+
Voronyj diagram
+
3D cities