Separate graphs of synonym and parent-child relations

[samm82]

From #192, I realized that I was using the word "graph" to mean "the image showing all the relations that I generate", but I now know I was misinterpreting this. "Graph" really means "a representation of a single relation", so each of my generated images really visualized two graphs on the same set of nodes. Because of this, we wondered if we should be displaying these relations separately.

During our meeting, I linked the (large) "graph" of all explicit relations between all explicit test approaches. Just now, I manually adapted the generated digraph code to create one image for each type of relation (again, still explicit, so this is still smaller than the set of relations/approaches that includes all not explicit data as we've previously discussed). I think taking a look at the generated explicit synonym graph helps defend why I think merging two relations into one graphical relation makes the most sense. 😂 We could remove all approaches that do not have a synonym relation, resulting in a nice small graph of synonyms to include in our thesis (perhaps alongside our definition of the relation?) but I'm not exactly sure what value that would bring. Thoughts?

[JacquesCarette]

Visualizations are meant for human enlightenment. So for the 'explicit synonym graph', this should only include the nodes involved in a non-trivial relationship. Some people are very visual, and so that illustration would be a compact representation of the information contained in that relation.

[samm82]

Alright, this looks a lot more useful and is automatically generated! 😁

[smiths]

Yes, this does look useful. 😄 If you put this figure in your thesis, you'll want to do some manual formatting so that you have more than one row of sub graphs.

[samm82]

Agreed; I'm also noticing that some of the "multi-synonyms" made it through, despite some of their synonyms being implicit (and therefore excluded. 😂 I'll also add a legend at some point as well, which would double as an example of how we color the relations (which was one of your critiques in #149).