Protocol: how the pool of sources was built

This file records every search, count and rule of the information-typing review. Shared rules: evidence README. Searches ran on 2026-10-29.

1. Why grey literature

Garousi 2019, Table 4, answered for information-typing: all seven questions get a yes.

# Question Answer for this review
1 Complex, not solvable from formal literature alone? yes: the main typed systems (DITA, Information Mapping) come from practice
2 Lack of volume or quality of evidence? yes: one controlled test of a typed format against untyped text was found, in a 1969 government report
3 Context important? yes: the domain (software, aerospace, schools) decides which types exist
4 Corroborate research with practice? yes
5 Challenge practice with research, or the reverse? yes: the DITA specification cites research this search could not find (position.md)
6 Would a synthesis serve both communities? yes
7 Large practitioner volume? yes: 802 firms are listed as using DITA (DITAWriter 2026)

2. Questions

Id Question
Q1 Which systems define a base unit and derive types from it, and what does the base hold?
Q2 Does a typed, predictable structure help readers find, understand or recall, measured?
Q3 Does knowing one type transfer to another type, or to a derived type?
Q4 What does practice adopt, with counts?
Q5 What does typing cost: rigidity, authoring burden, failed fit?
Q6 Do typed units help a language model as reader, kept apart from vendor claims?

3. Existing reviews (Garousi Guideline 2)

  • Searched first, by strings F10 (meta-analyses of text structure instruction), F29 (reviews of DITA and structured authoring) and F18 (component content management).
  • Found: 8 reviews. 3 meta-analyses read by abstract (one also through its thesis chapter), 1 review read by abstract, 3 reviews of component content management by title and venue only. 1 review of content strategy located only (Clark 2016, not used).
  • No systematic review of DITA, Information Mapping or topic types measured on readers was found. They are placed in review.md strand 0.

6. Stopping rule

  • Stated at the start of the formal search: effort bounded, N = 20 most cited per Semantic Scholar string, with saturation, k = 5 strings in a row that add no new included source.
  • Formal: evidence exhaustion for F3, F10, F23, F27, F28 (fewer than 10 records); effort bounded for the rest. The genre strings F12 to F15 and F21 reached saturation for reader-side evidence: they added writing-studies theory, not reader measures.
  • Grey web: first result page per string; saturation for G2.
  • Q/A: as above, ended after TU14.

7. Selection

  • Include: a source on how a unit of writing is given a type, how types derive from a base, or how readers or models use typed units. Also in English, and readable on the day at least by abstract or, for a book, by its publisher's page.
  • Exclude:
    • automatic genre classification and topic modelling (the words "topic" and "type" in another sense);
    • school writing interventions with no reading or typing claim;
    • reuse, maps, schemas and validation (document-model review) and nested display (progressive-disclosure review), passed to those reviews when met;
    • sources whose abstract could not be read and whose title does not settle relevance (kept as located only).

8. Quality of grey sources (Garousi Guideline 11, Table 7)

  • The 20 criteria and the threshold: evidence README.
  • Outlet tier: books, theses, government reports and standards = tier 1 (1); company and project pages, Q/A threads = tier 2 (0.5); blogs and informal lists = tier 3 (0).
  • A part not seen scores 0; a book known only from its publisher's page scores low for that reason, and says so.
  • Threshold: 10 of 20. Below it a source is an example of practice only, never support for a claim.
  • Scores, with subtotals, are in sources.md.

9. Extraction and synthesis

  • Extracted per source: kind, year, the claim, the evidence offered, any number (marked "(secondary)" if not read in the primary), and which question it answers.
  • Synthesis is narrative, by question Q1 to Q6; every claim traces to a row of sources.md.

10. Limits

  • One first rater for information-typing; the second rater's blind sample is reported in the evidence README.
  • Shared limits: evidence README. Here: two arXiv strings were refused.
  • Most formal sources were read by abstract only, several by title only; sources.md says which. Publisher pages of Elsevier, Wiley, IEEE and APA were not readable without access.
  • No archived copy was made of the practice pages; three of them (DITAWriter, Information Mapping, a blog) may move.
  • This search did not find a reader study of a DITA topic type; that is not proof none exists.