See Rieks' article on a Real Life Example.
- Neil Thomson said "terms are a set of objects in a graph". He explained the problem of multiple usages of the same term — and that the solution is qualification of the terms.
- Rieks Joosten explained that this is the purpose of his on Real Life Example article.
- His key point it that it is not the term that matters, it is the meaning of how it is distinguishes between what is and is not identified by the term.
- There is no "right" definition, only a consensus on how to use the term.
- As an example, Rieks recommends that the Technology Architecture TF should not focus on agreeing on one definition for "integrity", rather to focus on what they actually need to distinguish. Once those differences in meaning are agreed, then they can agree on the terms to assign to them. Then they need to commit to use those terms for those meanings. That consistent usage within the context will start being reinforced by everyone.
- The two principles are:
- First try to reach agreement on the meaning. Only once the meaning is agreed, then decide what term works best within the terms community for that meaning.
- Commit to using the term assigned to that meaning. If the meaning turns out to be incorrect or not useful, then repeat the process.
- Neil Thomson said that everyone wants to be economical in speech, so that's part of the problem.
- Rieks Joosten explained that this is the primary reason for mental models, because they will help draw both the distinctions and the relationships between the terms, which makes their meanings much clearer for everyone.
- We discussed the TATF example of the term "integrity". Rieks suggested applying his proposed process to determine the "five definitions" that have been used. If there is agreement on those five meanings, then we will assign five distinct terms to those meanings.
ACTION: Drummond Reed and Neil Thomson to coordinate on an action plan for the Technology Architecture TF terms wiki and glossary.
- Nicky Hickman pointed out that "terms are also language" and thus can have different meanings. She used the example that the word "interesting" has different meanings in America and Europe. So it is a standard "wetware feature of terminology" that we can't engineer out of it.
- She also pointed out that sometimes it is okay to let there be some relatively small ambiguity about terms that have roughly the same meaning but have particular "politics" associated with them. Agreeing to more precise meaning becomes more important in technical documents.
- She used the example of "digital identity ecosystem" and said that they agreed that it was roughly analogous to "digital trust ecosystem".
- Rieks acknowledged everything that Nicky said and doesn't feel it is a contradiction because:
- You don't have to have agreement on a detailed meaning if that's not needed. His example is the term "white". People from the northern climates can have 30+ shades of meaning for "white", but the root term is still sufficient for many usages.
- The same can apply to a term like "identity" when a precise definition is not needed.
- Another point is that having more than one term that refers to the same meaning is fine — as long as that meaning is agreed.
- Nicky summarized by saying "the golden rule is meaning" — as long as the meaning is conveyed, the goal is achieved.
- Neil gave the example of the term "trust" — it is very broad.
- Rieks applied his methodology by then asked what criterion makes the distinction that is actually needed in the work being done.
- Neil advocated using a qualified term so that the audience understands the relationship to the root term.