Introduce "Flaw" as a catch-all for what we've called "Discrepancies"

[samm82]

In #137, @JacquesCarette came up with the term "flaw" to refer to any instance of something being "wrong" with software testing terminology. We clarified our path forward for defining this set of terms in #151.

• Use "flaw" in TeX files
• Use "flaw" in LaTeX macros
• Use "flaw" in file names
• Use "flaw" in Python variable names
• Use "flaw" in Tex labels
• Move Table 4.3 to Terminology and split into its orthogonal parts
  • Move definitions from Flaws section to Terminology section, with pointers back where relevant
    I think I've done this; we'll come back to this when reviewing the Flaws section to see if anything there should be moved.
  • Define each "semantic" flaw
    • Rename "terminology" flaws to "label" flaws (Thesis Meeting | Mar 3, 2024 - 2:30pm - Teams #156)
    • Rename "citation" flaws to "traceability flaws" (Thesis Meeting | Mar 3, 2024 - 2:30pm - Teams #156)
  • Define each "syntactic" flaw
• Define "singular" versions of kinds of flaws (so that we can say "definition flaws" with the hyperlinked macro/flaw ref without having to say "flaws with definitions")
• Use "inconsistency" as a synonym for "flaw" when comparing multiple sources instead of "discrepancy" (Proposed Taxonomy of Flaw-related terms #155, Thesis Meeting | Mar 3, 2024 - 2:30pm - Teams #156)
• Remove "in-depth" research into why "flaw" is a good term to use (maybe as a footnote). (Thesis Meeting | Mar 3, 2024 - 2:30pm - Teams #156)
• Explain the rationale behind the term "flaw" by reverse engineering what it means for something to be "wrong"
  This will also give us more complete criteria when making recommendations. Currently, the 'manifestations' seem to provide good anti-criteria…
• Include a brief summary of this terminology, likely in the introductory chapter, emphasizing that "flaw" is the general term

[samm82]

I've started "trying out" this term when writing new content, and noticed that talking about a "flaw between two sources" seems a little awkward. Do you think this phrase is alright, or could we also use the term "discrepancy" when explicitly comparing two sources of information? For clarity, these would also be considered flaws, and I would outline this explicitly. @JacquesCarette @smiths

[JacquesCarette]

You are right that "flaw between two sources" sounds awkward. I'd be tempted use "inconsistency" as a replacement in those cases (but willing to hear arguments that this is also, ahem, flawed.). Google's AI and thesaurus.com find interesting synonyms (I googled 'flaw synonym').

[samm82]

I think I'll start by introducing "flaw" for:

1. catch-call situations
2. self-contained flaws

then see what remain in case that influences our decision (but yes, "inconsistency" sounds good in the absence of any better alternatives!)

[samm82]

So far, I've found the following cases where a "discrepancy" has an implicit comparison:

• These kinds of discrepancies [differences between "error", "fault", and "failure"] can lead to miscommunications…
• Discrepancies between groups are not double counted; this means the following line adds discrepancies between X and Z and between Y and Z, without counting the discrepancy between X and Z twice.
• Since the IEEE documents are written by the same standards organizations (ISO/IEC and IEEE), they are counted as a discrepancy between documents by the same author(s) in Figure 4.1.
• The rest increment their corresponding count in Figure 4.1 and Tables 4.1 and 4.2 by only one:
  • an explicit discrepancy between documents by ISO/IEC and IEEE,
  • an implicit discrepancy ^{This might just be a flaw; don't get caught up on this one!} within a single document, and
  • an implicit discrepancy between a paper and a standard.

I think the word "inconsistency" works for all of these, for the same reason I think "flaw" works for the general case. I originally liked "conflict", but that's too similar to "contradiction", which I define more formally, so "inconsistency" sounds good to me. However, upon more thinking, I think "discrepancy" is still the best pick for this distinction, as I explain in #151.