| name | haipipe-board-page-for-venue |
| description | The VARIANT contract for a Board's numbered QBv venue Pages, one Page per submission target such as a journal, agency, or patent office. It loads haipipe-board-page for the base frame and adds three figures in fixed order, venue-indexed Content divisions, the rule that the desk outranks the pack, five Files groups, provenance for every desk fact, and visible open rows for facts the pack cannot supply. Use when writing or fixing a venue Page, when a new outlet gets a Page, when the pack and desk disagree, or when a venue Page states a number with no source. Trigger: venue page, QBv, outlet page, journal page, desk, venue pack, playbook, submission rules, page limit, Venue-Structure, Submission-Rules, exemplars, section kinds, /haipipe-board-page-for-venue. |
| metadata | {"version":"0.1.2","last_updated":"2026-08-04","summary":"Now lives under page-types/ and composes the venue-page contract with the current Page Phase."} |
/haipipe-board-page-for-venue · a venue page records a desk it cannot argue with
LOAD haipipe-board-page FIRST. It owns the base: the sections and their fixed order, the five rows that define each one, the title rule, the Opening split, the numbering, and the evaluation contract.
This file adds only what a venue page needs and an ordinary decision page does not.
It never repeats a base rule, because a copied rule goes a night out of date while the contract moves.
After resolving this Page Type, load the current contract from page-phases/; this file determines the persistent venue shape, not whether the current work is DRAFT, PROBE, REVISE, or CHECK.
The kind this variant covers: one page per submission TARGET, and nothing above it.
kind filename subject closes when
────────────────────────────────────────────────────────────────────────────
Venue QBv<n>-<slug>.md ONE desk: a journal, a its Aims are met
funder, a patent office like any Q page
QBv1-misq.md is the reference implementation. Read it before writing a new one.
There is no pack-head page above the outlets (JL 260802): four outlets in one pack get four pages, and what they share is stated on each page that needs it, never in a fifth page nobody opens.
🏛 What makes a venue page different
A normal Q page settles something the team gets to decide. A venue page settles nothing. Its subject is a desk outside this repo that publishes its own rules, changes them without telling anyone, and rejects papers that ignore them. The page's job is to make that desk legible before a paper is written for it, not to have an opinion about it.
Three consequences, and every rule below comes from one of them:
1 the subject is EXTERNAL → every fact carries where it came from and when
2 there are TWO sources → the desk outranks the pack, and the gap is written down
3 a paper is BUILT from it → the page owes structure and mechanics, not just taste
What a venue page does NOT own: which venue this paper picks. That decision lives in the paper board's Opening concern (QB1 on the paper board). This group is the catalog that decision reads.
📖 A venue page is a REFERENCE, not a rulebook
The principle (JL 260803), and it governs everything below. A venue page is read while writing a paper, and almost everything on it is a suggestion: the arcs, the budgets, the moves, the anti-patterns are all measurements of what published papers at that desk actually did. A drafter may depart from any of them on purpose, and doing so is off-pattern rather than wrong.
One part of the page binds. The desk's own published rules, which live in the Submission-Rules figure and the Authority group, are enforced by the desk itself: break one and the manuscript is returned unreviewed. Nothing else on the page has that force.
📖 REFERENCE arcs · budgets · shapes · moves · the pack's refusals ·
format values · exemplar language
↳ measured from papers · departing is a choice
⚖️ BINDING category cap · manuscript format · reference style ·
anonymity · required disclosures · the portal
↳ published by the desk · breaking one costs the submission
Why it has to be written down. A page full of measured numbers reads as a specification. QBv1 printed 120-160 words, do NOT exceed ~185 for the abstract, which is the pack's measurement of eight papers, while the desk publishes no abstract cap at all. Read as a rule, it makes a drafter write to the middle of a distribution instead of to the paper. Found by JL 260803, who ruled ~250 words fine.
So every length says whose it is, and the page says it in the row itself:
✅ 120-160w observed, no desk cap · ~250w ruled acceptable JL 260803
✅ 55 pp Research Article · the DESK's cap, counting everything
🚫 120-160 words, do NOT exceed ~185 ← reads as a rule, is not one
A binding rule also has a WHEN, and it is not always submission. Nature Communications requires no particular structure or format at first submission and says style and length "will not directly influence consideration"; its caps are enforced at REVISION. A page that files those caps beside JAMA's tells a drafter to spend weeks on a gate that is not there yet, and a page that omits them lets a revision arrive over the limit. So a binding row carries the moment it bites:
⚖️ AT SUBMISSION the desk checks it before an editor reads · JAMA's 3,000 words
⏳ AT REVISION free at first submission, enforced later · Nature Communications
🎯 AT ACCEPTANCE priced or demanded once accepted · every Nature-family APC,
and npj's waiver, which must be REQUESTED at submission anyway
Write the pack's refusals as the pack's. the pack refuses more than ~160 words rather than do not exceed ~160 words, so the page never sounds like it is the one doing the refusing.
📚 The two sources, and which one wins
📦 THE PACK paper/venue/playbook-<family>/<OUTLET>/
taste.md · <OUTLET>-<section>/style.md + template.md · examples/
⚖️ READ and NEVER written by this plugin: the packs are their
own repository, jluo41/Venue-Paper, pinned as a submodule
🏛 THE DESK the venue's own published instructions, on its own site
submission guidelines · categories and lengths · policies ·
review process · the manuscript templates it publishes
⚔️ WHEN THEY DISAGREE the DESK wins, and the disagreement is WRITTEN DOWN on the
page, naming both readings. A correction never goes into
paper/venue/, because this plugin does not write there.
The disagreement is the most valuable thing on the page, so it is never quietly resolved. QBv1 records three of them: the pack observed "40-50 published pages" where the desk sets a 55-page submission ceiling, the pack records no reference style where the desk requires APA 7th, and the pack records none of the submission mechanics at all.
Provenance is stamped, and HOW it was read is part of it. A desk fact carries the date and the method, because these two are not equally strong:
✅ fetched and verified 260802
✅ re-checked 260803 through search summaries only, the site answers a direct
fetch with HTTP 403 · re-read before submitting
🚫 "MISQ requires APA 7th" ← true, and unciteable, and unmaintainable
🖼 Three figures, in this order, in ## Diagram
The base allows any number of figures and requires a caption line above each. A venue page carries these three, and a page missing one is incomplete rather than short:
① DESK TASTE what COUNTS as the contribution · what is DESK-REJECTED ·
the desk's own test, quoted · the exemplar count
answers: would this desk even look at my paper?
② VENUE-STRUCTURE the sections in the venue's READING ORDER, each with its
budget and what it owes · the one ceiling over all of them ·
whether a kind is a floor a real paper may exceed
answers: what am I actually writing?
③ SUBMISSION-RULES category and cap · manuscript format · reference style ·
the submission system · anonymity · required disclosures ·
the odds, the clock and the money · an open row for what is
NOT on record · the desk's own URLs
answers: what does the portal demand on the day I upload,
and what am I signing up for by choosing this desk?
When the desk publishes a TOTAL, add the pack's parts up and print the sum against it. This is a required row in Venue-Structure, not an optional nicety, because it went 3 for 3 the first time anyone checked:
JAMA IM desk 3,000 w body · pack floors sum to 3,350 · over before you start
PNAS desk ~4,000 w · pack floors sum to ~6,970 · 74% over
Diabetes Care desk 4,000 w body · pack floors sum to 3,950 · the cap IS the floor
MISQ desk 55 pp counting everything · no per-section desk limit at all
The arithmetic is systematic rather than unlucky. A pack measures PUBLISHED papers section by section, and a published page is not a submission budget, so the parts were never fitted to the whole. Neither source is wrong and the page resolves nothing: it prints both, names the cap as the binding one, and opens an Aim on how the body gets allocated. A drafter who writes to the middle of every per-section budget is over the cap at every desk that has one.
The odds and the clock belong on the page. Acceptance rate, how many review rounds a paper takes, how long a cycle runs, whether submission or open access costs anything: these decide whether a desk is worth a year, and none of them are in any pack. They are the desk's reported statistics rather than promises, and the row says so.
A venue page ends with the gate as a runnable list. The last Content division turns the binding figure into an ordered checklist run once on the final file, because a rule remembered while drafting is a rule half-applied. QBv1 runs seven steps: count the pages, pick the category, strip the author out, fix the references, move nothing outside the limit, write the disclosures, upload. The division also names the ONE step a finished paper cannot fix in an afternoon, which at every desk with a hard cap is the page count.
Embed the desk's links twice, on purpose. A row inside a fenced figure is plain text: the renderer runs esc() and the figure linker over a fence, never the inline markdown pass, so a URL written in a figure will not be clickable. Put the bare hosts in the figure so it stays readable when copied, and repeat them as real markdown links in a line directly under the fence.
Never put a bare URL alone on its own line inside ## Diagram. That is the Excalidraw canvas slot. A desk URL alone on a line is safe only because the renderer matches the excalidraw host specifically; do not rely on it, and keep desk links inline in prose.
🔢 A section division carries the venue's index
The rule (JL 260803, "I want to see the index"). A Content division that describes one section of the paper is named with the venue's own reading position in front of the section name:
### 4 · Sec-0-Abstract: one unstructured paragraph, question forward
▲ ▲
│ └── Sec-<n> from ZERO, so the index IS the S-Main page number ·
│ only the appendix takes a letter, Sec-A
└── the Content division number, which counts the judgment divisions too
### A4 · 📝 Sec-0-Abstract: one unstructured paragraph, question forward
▲ the emoji lives on the Aims and States group and on the division's
closing line. NOT on the division heading, however tempting the
symmetry: check.py strips it from a group name and not from a
division, so ten divisions produce twenty group-name-drift
warnings at once. Written, built, reverted, 260803.
The index counts from ZERO and carries the Sec- prefix (JL 260803, arrived at in three tries). A 1-based index sat one off the S-Main-<n> page forever, so every reader converted; 0-based fixes that, and Sec- is what keeps Sec-0-Abstract from reading as a typo where a bare 0-Abstract would. Letters were tried and dropped: a journal that letters its own appendix sections, as MISQ does with Appendix A and Table A1, gives a lettered section index two meanings at once. Only the appendix keeps a letter, Sec-A, precisely because it matches the desk's own lettering.
Sec-0-Abstract → S-Main-0 Sec-3-Methods → S-Main-3
Sec-1-Introduction → S-Main-1 Sec-4-Results → S-Main-4
Sec-2-Theory → S-Main-2 Sec-5-Discussion → S-Main-5 Sec-A-Appendix → S-Appendix-A
Sec-<i> = S-Main-<i> is a PROPERTY, not a law, and a page must check it rather than assume it. The identity holds only where the desk's reading order matches the order section-kinds.yml resolves, and three of the first four venue pages written to this contract broke it in a different way:
PNAS reading order is Significance ▸ Abstract ▸ … ▸ Methods LAST, while the
resolver has abstract=0, significance=1, methods=3 · the two columns
genuinely disagree, so QBv13 prints BOTH and explains why
JAMA NO no theory kind at this outlet, so Methods is Sec-2 where MISQ has Sec-3 ·
the index is the DESK's order, never a shared numbering across venues
JAMA IM a Research Letter is a whole article type with no reading position, so
QBv6 gave it Sec-L rather than claiming a slot in the sequence
When the two disagree, Sec-<n> follows the RESOLVER and the desk's printed order gets its own row. The fan-out split 2-2 on this before anyone ruled: QBv12 and QBv10 bound the index to section-kinds.yml, QBv9 and QBv13 bound it to the desk's reading order. Both printed both numbers, so nothing was lost either way, but a board cannot carry two conventions. The resolver wins for one reason: Sec-<n> exists to JOIN the S-Main page a paper actually writes, and an index that tracks the desk stops being a join key the moment the two differ. Describing the desk is Venue-Structure's job, and it has a column for it.
The desk's own section list can be longer than the resolver's, and that gap is a finding. section-kinds.yml declaring seven kinds is not evidence that the desk asks for seven blocks: MISQ expects Concluding Remarks that the resolver does not declare, Management Science requires a ≤250-word nontechnical executive summary that no pack folder covers, and JAMA Network Open publishes a Research Letter the resolver withholds. A draft built from the resolver alone reaches the portal with a required field empty. Print the block in Venue-Structure with the gap named, and open an Aim on it.
A section has three numbers, and only one pair agrees, so Venue-Structure prints all three: the venue index counts from 0 and the lifecycle page S-Main-<n> counts from 0, so those two LINE UP, which is the whole reason 0-based was chosen; the Content division is the one that differs, because it counts the judgment divisions ahead of it. State this in the figure; do not make a reader work it out.
📐 The division shape a section division repeats
Divisions §1 to §3 are the desk's judgment layer. Every division after them describes ONE section kind and repeats the same five parts, which is what makes the sections comparable across venues:
### <n> · <index>-<Kind>: <what this section is, in the desk's terms>
**Caption**: the one thing that is true of it here and not elsewhere.
```text fence: 📐 ARC · 📏 BUDGET · 🧱 SHAPE · 🔀 VARIANT or 🖼 DISPLAYS OWED
<emoji> Establishes ... one line, what this division settles
#### <n>.1 · The moves, as slots fill the shape; never lift the sentence
#### <n>.2 · What the pack refuses each a named anti-pattern, not a preference
#### <n>.3 · Format values words · citation density · value density · displays
#### <n>.4 · The language, in the papers' own words 5-6 attributed sentences, one move each
The judgment divisions are not fixed in number, and QBv1 runs three: what the desk buys and refuses, what arriving here costs, and which sibling outlet a paper leans to with the tie-break that pins it here.
Every number states its source inline. A word budget, a paragraph count, or a citation density may appear only with the line that records it: a style.md line number, an exemplar name, or the desk's own page. Never as the page's own claim.
A slot the sources cannot fill is PRINTED. 🔢 VALUE DENSITY not recorded by the pack is a finding; deleting the row is a silent gap. The same holds in Submission-Rules, which carries an open ❓ NOT ON RECORD YET row.
Say what the desk REFUSES, not what it prefers, because a preference does not decide a submission:
✅ a better classifier is not a MISQ paper
🚫 MISQ values theory highly
🏛 When the target is NOT a journal
A funder and a patent office are venue targets, and four of this contract's rules break on them. All four were found on 260803, by the agents that wrote QBv15-grant and QBv16-patent, and each is a real adaptation rather than a licence to skip the section.
1 Sec-<n> HAS NO RESOLVER section-kinds.yml declares ZERO kinds for grant and
for patent, so no S-Main page exists and the index
cannot be "the S-Main number"
→ grant used Row-<i>-<AGENCY>, the pack's own unit
→ patent read its index off 37 CFR 1.77(b), with
HOLES: conditional items take a figure row and no
division, so the index is not contiguous
2 ONE TARGET, MANY DESKS playbook-grant covers 8 agencies, playbook-patent
covers CNIPA + USPTO + EPO, each with its own
prescribed order
→ Venue-Structure becomes a MATRIX, one lane per
agency; or PIN one, say so in three places, and
keep the others' deltas as their own divisions
3 "PACK OBSERVATION" IS EMPTY the label means measured from published work, and
these packs hold ZERO funded proposals and ZERO
filings
→ use AGENCY RULE vs PACK RECORD, and say on the
page why the measured tier is missing
4 A RULE NEEDS ITS CYCLE a journal's guidelines stand until changed; a
funder's expire every round and an office's fees
are dated
→ provenance says when it was READ · a non-journal
row also says which CYCLE it binds
The unfixable step is not the page count. At a journal it usually is. At a funder it is CHOOSING THE AGENCY, because a 15-page NSF description, a 5-page ERC Part I and a 7-page ARC description are three documents rather than three formats of one. At a patent office it is what the specification failed to disclose, because 35 U.S.C. 132(a) bars new matter after the filing date. Verify the claim at source before writing it, as QBv16 did.
A blueprint-only pack has no exemplar language, and the page says so. The per-section shape (n.1 moves, n.2 refusals, n.3 format values, n.4 the papers' own words) presumes measured work. Fill only what the pack or the office actually says, and never print four subsubsections with empty rows to satisfy the shape.
📎 Files: five groups, and two of them are this kind's own
The base offers a menu of action groups. A venue page takes these five, in this order:
⚙️ Engines what REGENERATES this page _tools/sync-exemplars.py
📋 Contracts THIS FILE, and the base it haipipe-board-page-for-venue ·
extends haipipe-board-page
📥 Input files the pack files this page READS taste.md · style.md · README.md ·
stages/section-kinds.yml
🔗 Authority what the DESK itself PUBLISHES, read directly and never through
the pack · opens with the provenance stamp · carries the desk's
links AND every place the desk contradicts the pack
📤 Generated what a tool WRITES into this page, between markers
Authority and Generated are this variant's additions; the other three are the base menu's own names. Both additions state an action, which is the base's test for a group name.
This contract goes in Contracts, never in Engines (JL 260803 asked which). The base's split is what decides it: an Engine is something you RUN and open to change behavior, while a Contract is what CARRIES a rule to other pages, and a loadable spec that never executes is named there. The row also has to say the link runs both ways: the page is this contract's reference implementation, so a rule changed on the page is changed in this file in the same pass, and a sibling outlet page reads this file rather than reading the reference page.
🤖 The generated spans
_tools/sync-exemplars.py on the paper board owns two marker blocks inside ## Files, and nothing else:
<!-- exemplars:begin --> … <!-- exemplars:end --> one row per exemplar PAPER,
keyed by filename stem
<!-- kinds:begin --> … <!-- kinds:end --> the S-page roster, resolved
from stages/section-kinds.yml
python3 <board>/_tools/sync-exemplars.py
python3 <board>/_tools/sync-exemplars.py --check
Never hand-edit between the markers: the next run overwrites it. The blocks are replaced by a marker regex, so they may sit under any heading, but keep the kinds block after the exemplars block. Run --check before calling a page finished; a stale block is a count that disagrees with the folder.
🩺 state: on a venue page
The page keeps the base's four values, and the readable note answers one question: how much of this desk is actually recorded, and how much is still prose nobody reads?
✅ 🟡 PARTIAL · 15 exemplars · 7 sections · taste ✓ · the one-sentence test is
unread by any skill
🚫 🟡 PARTIAL ← says nothing a reader can check
🚫 🟡 PARTIAL · pack looks good ← a mood, not a count
The desk being well documented is not the same as the desk being wired in. A page whose facts are complete and whose Aims are all ⬜ is honest, and common: the pack is a library until a lifecycle stage reads it.
🎯 Aims and States
This variant does NOT override the base's Aim form. Ids, Done when, and one State row per Aim id all apply, unlike the skill-page variant, which drops them. Use ### A<n> · <emoji> <index>-<Kind>: <name> to mirror the Content division, and ### P for a target belonging to no single section, such as the submission mechanics or propagating a template to sibling pages.
An Aim on a venue page is almost always the same shape: something the pack records and nothing reads. Done when names the run that would prove it, not the reading.
✅ A4.1 · The abstract variant is chosen before drafting, since it decides
whether a number may appear.
Done when: an abstract draft records its variant and its measured word
count against the 120-160 budget.
🚫 A4.1 · Understand the abstract norms. ← no run can close this
Keep States honest against the folds. A Log line that says a question was settled and a State row that still says ⬜ Not started, and the pack cannot answer it is the drift this board exists to catch; it happened on QBv1 between 260802 and 260803.
🚪 The two verbs
📄 create a venue page
- Confirm the pack family and outlet folder exist under
paper/venue/. A missing file at either level is a missing ANSWER to record, not a reason to stop.
- Copy
QBv1-misq.md as the shape, never ref/page-template.md alone: the base template has no three-figure Diagram, no Authority group, and no section-division pattern.
- Take the next free
QBv<n> and register the page in board.md's ## Pages under QBv.
- Fill
§1-§3 from taste.md and the family README.md, then one division per section kind from that kind's style.md and template.md.
- Fetch the DESK itself. Write
Submission-Rules and the Authority group from it, with the provenance stamp, and record every contradiction with the pack.
- Run
sync-exemplars.py, then build, check, and read the RENDER.
🔧 work on a venue page
- Read the whole page first, including the folds.
- Run the checker and clear the mechanical findings.
- Then check the three things no checker reaches: does every number name its source, is every desk fact stamped with how it was read, and does any State contradict a Log line.
- Re-read the desk when a fact is older than the last issue. A page limit or a reference style changes without a changelog.
python3 <toolkit>/skills/board/haipipe-board/cli/build.py <board-folder>
python3 <toolkit>/skills/board/haipipe-board/cli/check.py <board-folder> | grep '^QBv'
python3 <board-folder>/_tools/sync-exemplars.py --check
The template. template.md, beside this file. ONE template serves every page of this type: two pages of it differ in what they say, never in what shape they are, so nothing has to be resolved before writing one. Copy it, fill every <slot>, and delete each RULE comment as you satisfy it.
📂 Files
page-types/haipipe-board-page-for-venue/
├── SKILL.md this variant contract
└── CHANGELOG.md version history
Owns no scripts. The base frame is haipipe-board-page; the engine is haipipe-board; the generator for the marked spans is the paper board's _tools/sync-exemplars.py; the group's own law lives in that board's board.md under ### QBv · Delivery Venue; the reference implementation is QBv1-misq.md.
