Baseline
Step 5 of the review: the baseline the repository's method is compared with, and how. The review is finished because this page names it. Terms are in frame.
Shape: baseline
The maintainer's practice, written down for this review
- State the problem in one sentence, then find what the field calls it.
- Review surveys first, then the seminal work, then citations forward; search each term by name, acronym and author.
- Name the baseline before building. A review without one is not finished.
- Leave a trail in a folder: the brief, the notes, the output, the pointers.
Named baseline: B1
- B1 for the steps: the six activities of the design science research process in Peffers 2006, entered at activity 1.
- B1 for the trail: the project organization of Noble 2009 and Wilson 2017: an overview file, a folder per project, a dated log.
- Most adopted comparators: CRISP-DM phase 1 for success criteria (Chapman 2000) and the commitments of the Scrum guide (Schwaber 2020). They are practice, graded convention.
What the practice adds to B1 and what it lacks
| Step of the practice | In B1 | Difference |
|---|---|---|
| frame: the field's name | activity 1 defines the problem | the practice also looks for the field's name |
| review, surveys first | not an activity; its resource is knowledge of current solutions | the practice makes it a step with a search method |
| baseline named | implicit in objectives and evaluation | the practice makes it the exit test of the review |
| plan, execution, refinement | activities 3 and 4; evaluation may iterate | B1 has demonstration and communication, which the practice leaves to the genres |
| trail of four files | overview file, folder, dated log | the practice names the brief and the pointers |
How to compare
- Unit: one job, from a brief to its output, done twice on a paired task: once with the practice, once with B1 alone.
- Round trips: count the times the person asks for a change after the first delivery.
- Rework: count the files or sections rewritten after the first delivery.
- Resumption: give a second person only the folder and count the questions they ask before they can continue.
- Grade: a result from a handful of jobs is a case study, not a measurement.
- Status: not run. This page names the comparison; no data exist yet.
What would change the baseline
- A measured study of scope-first or baseline-first moves a claim up from convention.
- A better-evidenced process model replaces B1 for the steps.
- A guide to the trail with a test of resumption replaces B1 for the trail.