Navigating Contextual Complexity with Graph Visualization
Diagram-focused standards face the same challenges as older , document-based solutions . Both drawings and documents can introduce inconsistencies that are hard to detect because of the difficulty of thorough analysis and the limitations of humans to correctly integrate engineering context that is spread across many diagrams . [ 10 ] Manually maintained diagrams may fail to reflect important elements because an engineer didn ’ t include them , or the standard didn ’ t allow for them . [ 11 ] Producing , maintaining and reviewing these necessary diagrams consume considerable engineering time . Under this pressure , only the most critical diagrams are produced , and many useful diagrams won ’ t be maintained .
SysML v2 uses a graph-based approach to modeling a system . [ 12 ] Graph-based data structures and data-driven visualizations are an important tool in the evolution of model-based system engineering , since they can readily support the constructs demanded of the system modeling language . Graphs readily support direct navigation between model representations based on the underlying system data . Data-driven visualizations are not predicated upon the existence of the correct manually created diagram , or a diagram specification that includes the items of interest . [ 13 ]
Graph analysis and graph visualization are tools that may be employed to detect model inconsistencies . Graph structures support seamless navigation between abstraction layers and other defined contexts , enabling collaboration between end users of different disciplines .
Automatically generated graphs of engineered systems have the capacity to illustrate model inconsistencies because a query of the model data will return what is actually present in the model , not what the engineer manually placed in the diagram . Automated , data-driven visualization of system models have the potential to reveal direct connections , which might otherwise remain implicit because they were not included in the prescribed or manually maintained diagram . Conversely , they may reveal missing connections , by showing unconnected ports on elements that should be related . [ 9 ]
Dynamic and interactive graphs and automated graph-based layouts support human comprehension of increasingly abstract and complicated data landscapes . [ 14 ] Manually creating visualizations for each of these cases is prohibitively expensive . Data-driven graph-based diagram generation is likely to deliver meaningful productivity improvements throughout a system lifecycle .
4 USERS AND USE CASES
The definitions below align well with the persona ’ s identified by Li et . al . [ 4 ] Individual users will move between these modalities if they are both a designer and user of graph technology , or both an operator and analyst of systems in operation . We present examples of how these reference personas are reflected across participants in the practice of system engineering .
Journal of Innovation 77