State of the art
Step 3 of the review: for each question, the best evidence apart from the practice most adopted. They are two records, held by different sources. Terms are in frame.
Shape: state of the art
Best evidence
| Question | Holder | Method | Measured how |
|---|---|---|---|
| Q1 round trips: cost of late defects | Menzies 2017 | analysis of 171 projects, 2006 to 2014 (secondary) | found no delayed-issue effect |
| Q1 round trips: a taught personal process | Prechelt 2001b | controlled experiment | positive effects, smaller than proponents assume |
| Q1 round trips: project success and a method | Serrador 2015 | 1,002 projects (secondary) | correlation |
| Q1 round trips: success figures | Eveleens 2010 | 5,457 forecasts of 1,211 projects (secondary) | the Chaos figures did not reflect the data |
| Q1 round trips: heuristics | Hembree 1992 | meta-analysis of 487 reports (secondary) | effect tied to trained teachers |
| Q2 topics: outlines | Kellogg 1988 | experiments (secondary) | judged quality of texts |
| Q3 trail | none found | none | none |
| Q4 boundary | none found | none | none |
Most adopted practice
| Question | Holder | Evidence of adoption |
|---|---|---|
| Steps from problem to result, in research | Peffers 2006, with its journal version Peffers 2007 | the paper describes 2 case studies; usage counts were not taken in this pass |
| Steps, in data work | Chapman 2000 | a published consortium guide; no count taken |
| Goals and done, in product work | Schwaber 2020 | a guide with a version history; no count taken |
| The trail | Noble 2009, Wilson 2017 | workshops reaching over 11,000 people (secondary) |
Upper and lower bounds
- Upper bounds, what is achieved: no result shows a cut in round trips from any step written first.
- Lower bounds, what is ruled out: a constant 100-times cost of late defects is not supported as a constant (Menzies 2017, secondary).
- The 40 to 50 percent share of avoidable rework (Boehm 2001) is an expert report, not a measured bound.
What the record lacks
- A comparison of projects with and without a written scope or a named baseline.
- A study of whether a set of files lets a second person resume.
- A study of a plan written as a structured text.