# Editorial guideline — what this project is, and how everything in it should be written

This governs every commentary, summary, article, book note and page on the site. Read it before
writing anything for publication.

## 1. What this is

**An encyclopedia of Pakistan's provincial textbook boards** — what each board teaches, how the
provinces differ, and what the differences imply for the generations being taught. Four boards,
classes 6–10, plus a historical layer reaching back to 1949.

Three things follow from calling it an encyclopedia rather than a report:

- **Completeness is the goal.** Every book the boards publish in scope should be catalogued, whether
  or not it is interesting, and whether or not it has been scored. A gap is recorded as a gap.
- **The reader should never have to hunt.** Every claim reaches its source in one click; every book
  record carries its provenance and a route to the book itself.
- **It outlives any single argument.** Findings will be revised. The catalogue, the citations and the
  method should remain useful to someone who disagrees with every conclusion in it.

## 2. The reader

Assume four, and write for all of them at once:

1. **A parent or teacher** who wants to know what their child is being taught, and has no patience for
   methodology.
2. **A journalist** who needs one quotable, verifiable sentence and the page it came from.
3. **A policymaker or board official** who will look for the error that lets them dismiss the whole
   thing.
4. **A researcher** who wants the data, the rubric and the limits, and will check them.

Reader 3 sets the standard. Anything that cannot survive a hostile, well-informed reading should not
be published.

## 3. How to write a commentary

- **Lead with what the books say, not with what it means.** The evidence carries the argument; the
  interpretation follows it and is labelled as interpretation.
- **Attach a citation to every claim** — board, book, class, edition, page. Page conventions differ
  (printed folio vs PDF index); state which is being used.
- **Quote only English-medium sources.** Urdu-medium material is theme-grade: describe structure and
  topics, never wording. Say "not established", never "absent", for an Urdu absence.
- **Give counter-evidence equal prominence, in the same section.** Not a footnote, not a final
  paragraph. Where a board contradicts itself, show both sides.
- **Name the vintage.** Four curriculum generations are in print. Where compared books differ in
  generation, say so and say which way it cuts. Where they cannot be matched, mark it confounded.
- **Do not claim effects on belief.** The corpus shows what is taught, repeated, sequenced and
  examined. It does not show what any child concludes. Say what the method can support, and say
  plainly what it cannot.
- **Prefer the specific to the sweeping.** "A 2022 Punjab geography book tells eleven-year-olds that
  people of the Congo are uncivilised, on printed page 36" beats any general statement about bias.

## 4. On implications — the "what you sow" question

The project's second half is the forward-looking one: what these curricula imply for the people
formed by them. It is legitimate and it is the reason most readers will come. It is also where
overclaiming is easiest.

Rules for it:

- Argue from **structure**, not from psychology: what is repeated across subjects and years, what
  sequence a pupil meets it in, and **what they are required to reproduce in an examination.**
- Assessment demand is the strongest available instrument, because an examinable proposition is
  compelled rather than merely offered. Use it where it exists, and say when it does not.
- Where an argument depends on an unobserved step — classroom practice, actual exam papers, what
  pupils believe — name that step as unobserved rather than assuming it.
- Say what would falsify the claim.

## 5. Sources, copyright and what may be hosted

The textbooks carry explicit reservations of rights — PCTB: *"All rights are reserved… No part of
this textbook can be copied, translated, reproduced"*; STBB: *"ALL RIGHTS RESERVED WITH THE SINDH
TEXTBOOK BOARD, JAMSHORO."* Free distribution to pupils is not a licence to republish. So:

**Host in full:**
- Everything this project wrote — analyses, articles, the rubric, the dataset, the scripts.
- Documents published for free distribution by their authors, where that is verifiable.

**Do not host in full:** the textbook PDFs themselves.

**Instead, for every book:** a complete bibliographic record (board, class, subject, medium, edition,
approval reference, page count), the analysis, the evidence with page citations, and **a direct link
to the source we obtained it from**, credited. The reader clicks once. That is what a good
encyclopedia does, and it is what a library catalogue does.

**Excerpt freely for criticism and analysis.** Quoting a passage to examine it, and reproducing a
page image as evidence for a specific claim, is ordinary scholarly practice and is how the argument
is made. Excerpt as much as the argument needs and no more; always with citation.

**Credit every source**, including the mirror sites, which are doing real public-access work even
where their metadata is unreliable.

**If a board asks for something to be removed**, remove it and say publicly that it was removed and
by whom. That record is itself part of the encyclopedia.

## 6. Standing accuracy rules

- Re-verify every quotation by exact string match against source before publishing.
- Where two of our own documents disagree, resolve it or publish both readings and mark it
  unresolved. Do not silently pick one.
- Correct errors in place, visibly, and keep the correction. This project has already corrected
  itself on the native-text count, Punjab's Urdu 9 vintage, Punjab's relocation (not deletion) of
  Muslim-scientist content, and a mis-set OCR language model. Those corrections stay on the record.
- A negative result is a result. Publish it with the same prominence as a positive one.