Visualizing Relationships among Components Using Grouping Information

• US 20100063785A1
• Filed: 09/11/2008
• Published: 03/11/2010
• Est. Priority Date: 09/11/2008
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Pich, Christian, Robertson, George G., Nachmanson, Lev B.
• Granted Patent: US 8,499,284 B2
• Time in Patent Office: Days
• US Class Current: 703/6
• CPC Class Codes: G06F 18/2137 based on criteria of topology preservation, e.g. multidimensional scaling or self-organising maps
• CPC Class Codes: G06F 18/40 Software arrangements specially adapted for pattern recognition, e.g. user interfaces or toolboxes therefor
• CPC Class Codes: G06T 11/206 Drawing of charts or graphs

Abstract

A graph processing module is described for visualizing relationships among components, such as software components within a software system. The graph processing module generates group results based on grouping information which identifies groups of individual components. The graph processing module uses the group results to bias component-level analysis that it performs on the components. In one case, the component-level analysis can involve similarity analysis (e.g., using multidimensional scaling) to assess the similarity among components and importance analysis (e.g., using a network page ranking algorithm) to assess the relative importance of the components. The results of the component-level analysis can be presented for visualization. The biasing effects of the grouping results provide insight into the relationships among the components.

Claims

1. A method for visualizing relationships among components, comprising:

receiving graph information associated with a plurality of components;

receiving grouping information that identifies groups within the plurality of components;

generating grouping results based on analysis of the graph information and the grouping information;

performing component-level analysis based on the graph information and the group results, the group results being used to bias the component-level analysis, the component-level analysis producing layout results; and

presenting a visualization of the layout results.

2. The method of claim 1, wherein the components correspond to software components in a software system.

3. The method of claim 2, wherein the graph information identifies the software components and the relationships among the software components.

4. The method of claim 2, wherein each group identified by the grouping information corresponds to a collection of software components that share at least one characteristic.

5. The method of claim 1, wherein the generating of grouping results and the performing of component-level analysis together operate to assess similarity among components.

6. The method of claim 5, wherein the generating of group results comprises:

coalescing nodes in a component-level graph G based on the grouping information, to form a group-level graph G_g;

determining distances d_ij between nodes in the component-level graph G; and

determining distances dg_ij between nodes in the group-level graph G_g.

7. The method of claim 6, wherein the performing of the component-level analysis comprises:

determining adjusted distances σ

_ij based on the distances d_ij obtained from the component-level graph G and the distances dg_ij obtained from the group-level graph G_g; and

performing similarity analysis using the adjusted distances σ

_ij.

8. The method of claim 7, wherein the similarity analysis is multidimensional scaling.

9. The method of claim 1, wherein the generating of grouping results and the performing of component-level analysis together operate to assess the importance of components.

10. The method of claim 9, wherein the generating of group results comprises:

coalescing nodes in a component-level graph G based on the grouping information to form a group-level graph G_g; and

determining an importance value for each node within the group-level graph G_g.

11. The method of claim 10, wherein the performing of the component-level analysis comprises determining an importance value for each node within a component-level graph G using the importance values computed for the group-level graph G_g as biasing influences.

12. The method of claim 11, wherein the importance values for the group-level graph G_g and the component-level graph G are calculated using a network page ranking algorithm.

13. A computer-readable medium for storing computer-readable instructions, the computer-readable instructions providing a graph processing module when executed by one or more processing devices, the computer-readable instructions comprising:

logic configured to perform similarity analysis to assess similarity among components, said logic for performing similarity analysis using grouping information to exert group-level influence on the similarity analysis;

logic configured to perform importance analysis to assess relative importance of the components, said logic for performing importance analysis using the grouping information to exert group-level influence on the importance analysis; and

logic configured to present a visualization of results provided by the similarity analysis and results provided by the importance analysis, at least one dimension of the visualization being associated with the results provided by the similarity analysis, and at least one other dimension of the visualization being associated with the results provided by the importance analysis.

14. The computer-readable medium of claim 13, wherein the components correspond to software components in a software system.

15. The computer-readable medium of claim 13, wherein at least one biasing parameter ω

controls an extent to which the grouping information influences the similarity analysis and the importance analysis.

16. A graph processing module for visualizing relationships among components of a software system, comprising:

a similarity analysis module configured to assess similarity among the components, said similarity analysis module comprising;

logic configured to coalesce nodes in a component-level graph G based on grouping information, to form a group-level graph G_g;

logic configured to determine distances d_ij between nodes in the component-level graph G;

logic configured to determine distances dg_ij between nodes in the group-level graph G_g;

logic configured to determine adjusted distances σ

_ij based on the distances d_ij obtained from the component-level graph G and the distances dg_ij obtained from the group-level graph G_g; and

logic configured to perform similarity analysis using the adjusted distances σ

_ij;

an importance analysis module configured to assess relative importance of the components, said importance analysis module comprising;

logic configured to generate importance values associated with nodes within the group-level graph G_g; and

logic configured to perform importance analysis for components within the component-level graph G, as biased by the importance values generated for the nodes within the group-level graph G_g; and

a graph layout module configured to provide a visualization of results provided by the similarity analysis module and the importance analysis module.

17. The graph processing module of claim 16, wherein the similarity analysis performed by the similarity analysis module uses multidimensional scaling.

18. The graph processing module of claim 16, wherein the importance analysis performed by the importance analysis module uses a network page ranking algorithm.

19. The graph processing module of claim 16, wherein at least one biasing parameter ω

controls an extent to which the grouping information influences the similarity analysis and the importance analysis.

20. The graph processing module of claim 19, wherein an adjustment to said at least one biasing parameter ω

controls an affinity of components within a group to other members of the group, said affinity competing with group-agnostic dependencies among individual components.