Use Cases

[richardmarcus]

A collection of important use cases, where TGView (3D) can support the workflows of scientists and others interested in any kind of graphs

• Accessing Information
  Scientific knowledge can be visualized as a graph compared to clicking through Wikipedia
  Links: further information can be accessed (web page or short definitions)
  MPDs: e.g simulation visualizations, etc

• Graph Editor
  MPDs: can be manually placed in a way that makes the underlying concepts easy to understand

• Result of Computation
  (Argumentation graphs?) Max' work can be visualized and different results can be generated by selecting another subset of nodes from the graph

• Interactive Graph Viewing
  Pretty pictures and smooth interaction

• Debugging
  Dependency graphs: code dependencies, error tracing can be supported by highlighting/tracking in the graph

• Effective layouting of directed/hierarchical graphs
  common layouting algorithms do not work well with those graphs

• General Framework for Implementing Graph Interactions
  Because of the wide range of basic interactions the tool can be seen as framework for further unknown use cases

• Representing Information
  MPDs: difficult concepts can be visualized by a graph for better understanding

• Discovering knowledge
  Visualizing the relations and structure of known facts can lead to new insights

[richardmarcus]

https://neo4j.com/
and surrounding software seem very interesting in this context

[richardmarcus]

also keep an eye on this
UniFormal/MMT#415

[rappatoni]

Another use case:

• Identifying Domains:
  Use the clustering algorithm to identify domains and domain membership.

When constructing theory graphs, it is not always obvious to which domain a theory should belong. For example, the Lebesgue Measure is important in Analysis for the Lebesgue Integral but also features in Probability Theory. In addition, one might argue it should be part of a separate domain altogether, Measure Theory.

Making these domain membership decisions can be quite difficult. Clustering could help with that. The resulting domain landscape should then be passed back to MMT and be integrated as meta-data on the archive (this pertains to the required functions mentioned in UniFormal/MMT#415).

[rappatoni]

And another one:

• Visualize Type Checking
  In the theory graph use color coding to distinguish the theories of the graph that currently type-check (green) vs the ones that don't.

[rappatoni]

Another use case:

Subsume the functionality of this tool:

https://tikzcd.yichuanshen.de/#
https://github.com/yishn/tikzcd-editor

That is, "draw" a theory graph in TGView and output tikz-code that can then be further edited in a TeX-editor.