The use of triangulation as an analogy here seems to suggest the choice of use-cases is critical to minimizing the number needed. In some cases I would think that selection might be rather obvious but not sure that would be generally true.
The obvious (I would think) goal then would be to seek use-cases that exhibit a high degree of “independence” from one another (much like a regression test should use) rather than “highly correlated” use-cases. But use-cases are not simple variables so may be difficult to assess from that perspective.
Do you have some thoughts on that selection process
That’s exactly right. I usually look for examples/use-cases which feel as qualitatively different from each other as possible, while still all feeling like they involve the same phenomenon we’re trying to deconfuse.
The use of triangulation as an analogy here seems to suggest the choice of use-cases is critical to minimizing the number needed. In some cases I would think that selection might be rather obvious but not sure that would be generally true.
The obvious (I would think) goal then would be to seek use-cases that exhibit a high degree of “independence” from one another (much like a regression test should use) rather than “highly correlated” use-cases. But use-cases are not simple variables so may be difficult to assess from that perspective.
Do you have some thoughts on that selection process
That’s exactly right. I usually look for examples/use-cases which feel as qualitatively different from each other as possible, while still all feeling like they involve the same phenomenon we’re trying to deconfuse.