Thesis Meeting | July 21, 2024 - 2pm - Teams

[samm82]

Updates

I've been chipping away at the Methodology/Tools chapters, and some chapters are almost ready for review:

1. Once How to improve/shrink example of recording test approach information? #198 is finished (see below) I'll re-assign review of the Methodology chapter to @JacquesCarette as per Dr. Carette's review of Methodology section #190.
2. I've been waiting on doing a final read-through of the Terminology chapter before assigning it to review to hopefully get some clarity on the below topics first; once I'm happy with it, I'll re-assign it to @smiths for (hopefully) its final review.

Questions

1. Follow-up on How to improve/shrink example of recording test approach information? #198 - really boils down to "how should I apply the flowchart idea to how we record software qualities?"
2. Follow-up on Separate graphs of synonym and parent-child relations #194 - essentially, is including an example parent-child visualization (which also demonstrates implicitness) and all the relevant synonym relations in our glossary (which also demonstrates "source tier colouring") sufficient?
3. Relatedly, if we don't include any examples of a focused visualization of relations between static approaches in our thesis, do we need an example of how we generate those visualizations? I think omitting it for brevity and clarity is a good idea. Relevant commit: d40bfde
4. (Part of Improve Tools section (mainly based on Dr. Smith's feedback) #191) What's a better way to explain what the second group of macros represent with less repetition? This is what I currently have; originally I just had "before/after process in Section 3.2.2":

[samm82]

Answers to Questions

1. I should make a flowchart for how we record software qualities (it's a more intuitive way to capture information vs. a paragraph, and if it's clear enough, I maybe be able to omit the example entirely!) (Obsolete; see How to improve/shrink example of recording test approach information? #198)
2. (Covered in Re-introduce "mock" example of synonym glossary + visualization #208) Main takeaway: I should use the mock example of a synonym glossary/visualization, moving the generated flawed synonyms to Observed Flaws and the full list of various kinds of these flaws to the flaw appendix (with appropriate points).
   Minor takeaways:
   1. I should replace 0000, etc. with actual years for clarity
   2. I should maybe include the actual CSV file for example glossary entries in an appendix? Then we can keep our "pretty" visualization (and make it even prettier by using an asterisk or something to indicate that D is implicit in the parent example) while still providing potential users of these tools with the information they need to use them.
   3. I should reframe the parent relation in legends by, for example, displaying "X ---> Y" and labelling the arrow as "one of X's parents is Y". I might do the same with synonym relations for consistency.
   4. My mock example of parent relations has an explicit relation to an implicit term. I included this to show that omitting an approach also omits all of its relations from the visualization, but this is definitely an edge case and I should see if this ever happens in my data (as a low priority activity).
3. What I did is good! 🥳
4. I should refer to these macros by the step they take place in: we had discussed "after step 3" and "after step 4", but simply doing "before/after step 4" might articulate the same information cleaner and be consistent with the current naming convention (i.e., less potentially traceability issues).

[samm82]

@JacquesCarette, it's been hard to find an example of child-parent graphs with a legend, but I've found some hierarchical graphs that label "child nodes" and "parent nodes" (leaving the relation between them implicit) on GeeksforGeeks, Medium, and CodeChef. Examples on Graphviz's website that match my visualization format do the same, both for "the 'family tree' of a fraternity and sorority" and "the UNIX Family 'Tree'". Even van Vliet (2000) seems to do this for directed relations: he defines the relation represented by a type of arrow (i.e., in a paragraph) then draws all the relations between nodes given in a graph as these arrows.

I guess I'm still struggling to see how modifying our legends this way would be helpful/less confusing, and I can't find any examples of it being done in the way you describe.