# Impose Desk

> A trim size is not a page size. The page is the trim plus the bleed on every side,
> so an A4 job with 3 mm of bleed is a 216 mm x 303 mm PDF — and a PDF is measured in POINTS, which are exactly 1/72 inch, so nothing
> metric is round: A4's trim is 595.276 pt x 841.89 pt and its page is 612.283 pt x 858.898 pt. **ANYTHING THAT FITS THAT TO A4 SCALES IT TO 97.22%**: the artwork shrinks, every element moves inward, the crop marks
> land 2.92 mm inside the trim, and the finished piece comes out 204.17 mm x 288.75 mm instead of 210 mm x 297 mm. **CREEP EATS THE BLEED**: 64 pages stitched on 0.1 mm stock is 16 sheets and 1.5 mm of creep — 50% of a 3 mm bleed. **AND HOW YOU BUTT THE PAGES DECIDES THE N-UP**: on SRA3 an A5 goes 4-up with shared bleeds and 2-up with separate ones.

## The one thing to know

**A TRIM SIZE IS NOT A PAGE SIZE.**

```text
point      = 25.4 / 72                                = 0.352778 mm, exactly
media box  = trim + 2 * bleed                          <- what the FILE has to be
marked box = trim + 2 * (bleed + offset + length)      <- what an IMPOSITION needs
fit scale  = min(target w / page w, target h / page h) <- the tighter axis wins
sheets     = pages / 4                                 <- one folded sheet, four pages
creep      = (sheets - 1) * caliper                    <- at the innermost spread
shingle    = creep / 2                                 <- half, not all of it
pitch      = trim, if the bleeds are shared
           = trim + 2 * bleed + gutter, if they are not
fits       = floor((usable + gutter) / pitch), per axis, rounded DOWN
printable  = (sheet w - 2 * margin) x (sheet h - grip - margin)
```

**A POINT IS A DEFINITION.** 72 to the inch and 25.4 mm to the inch, so a point is 0.352778 mm exactly — and every metric page size is therefore an untidy number of them.
A4's 595.276 pt x 841.89 pt is what a PDF's boxes actually contain, and every
rounding of it is an error somebody downstream inherits. The imperial sizes are the
same problem from the other end: US Letter is exactly 612 x 792 pt and 215.9 mm wide.

**AND "FIT TO PAGE" IS THE ENEMY.** The media box is bigger than the trim by two
bleeds, so anything offering to fit it to the trim scales the artwork down: on A4
with 3 mm that is 97.22%, the marks land 2.92 mm inside the trim rather than on it, and the finished piece is 5.83 mm narrow and 8.25 mm short. The fit is
governed by the TIGHTER axis, which is why a portrait page loses more height than
width for a bleed that is the same on all four sides. And it happens after the file
leaves you, which is why the number is worth knowing.

**CREEP EATS THE BLEED.** A saddle-stitched booklet is folded sheets nested inside
each other, so the inner fore edges are pushed outward by `(sheets − 1) × caliper`. 64 pages on 0.1 mm stock is 1.5 mm — 50% of a 3 mm bleed — and 96 pages on 0.15 mm stock is 3.45 mm, which is 115% of it. The content has to move
inward by up to HALF the creep, progressively from the middle, because the spreads
move in opposite directions from the centre of the book — which is what shingling
means. And caliper is not weight: two papers of the same grammage can differ by half
again in thickness.

**AND HOW YOU BUTT THE PAGES DECIDES THE N-UP, not how big they are.** Two pages
sharing a trim edge share one cut and one bleed, so the pitch is the trim size (148 mm for A5). Two with their own bleeds need a gutter of at
least twice the bleed, so the pitch is 222 mm. On SRA3's 310 mm x 435 mm of printable area that
is 4-up against 2-up — half the sheet and twice the paper, for a decision about where the cuts go.

**A press sheet is always smaller than the sheet**: the press grips one edge and
cannot print there, so 10 mm comes off the length once and 5 mm off each of the other three. SRA3's 320 mm x 450 mm is 310 mm x 435 mm to print on — 93.6% of what you bought.

## Every trim size, in both units

| Size | Trim | In points | With 3 mm of bleed | That page in points |
| --- | --- | --- | --- | --- |
| A3 | 297 mm x 420 mm | 841.89 pt x 1190.551 pt | 303 mm x 426 mm | 858.898 pt x 1207.559 pt |
| A4 | 210 mm x 297 mm | 595.276 pt x 841.89 pt | 216 mm x 303 mm | 612.283 pt x 858.898 pt |
| A5 | 148 mm x 210 mm | 419.528 pt x 595.276 pt | 154 mm x 216 mm | 436.535 pt x 612.283 pt |
| A6 | 105 mm x 148 mm | 297.638 pt x 419.528 pt | 111 mm x 154 mm | 314.646 pt x 436.535 pt |
| B2 | 500 mm x 707 mm | 1417.323 pt x 2004.094 pt | 506 mm x 713 mm | 1434.331 pt x 2021.102 pt |
| SRA3 | 320 mm x 450 mm | 907.087 pt x 1275.591 pt | 326 mm x 456 mm | 924.094 pt x 1292.598 pt |
| SRA2 | 450 mm x 640 mm | 1275.591 pt x 1814.173 pt | 456 mm x 646 mm | 1292.598 pt x 1831.181 pt |
| US Letter | 215.9 mm x 279.4 mm | 612 pt x 792 pt | 221.9 mm x 285.4 mm | 629.008 pt x 809.008 pt |
| US Legal | 215.9 mm x 355.6 mm | 612 pt x 1008 pt | 221.9 mm x 361.6 mm | 629.008 pt x 1025.008 pt |
| Tabloid | 279.4 mm x 431.8 mm | 792 pt x 1224 pt | 285.4 mm x 437.8 mm | 809.008 pt x 1241.008 pt |
| DL | 99 mm x 210 mm | 280.63 pt x 595.276 pt | 105 mm x 216 mm | 297.638 pt x 612.283 pt |
| 148 mm square | 148 mm x 148 mm | 419.528 pt x 419.528 pt | 154 mm x 154 mm | 436.535 pt x 436.535 pt |

**A point is exactly 1/72 inch**, so 25.4 mm to the inch makes a point 0.352778 mm — and no metric size is a round number of them. A4's 595.276 pt x 841.89 pt is the number a PDF's boxes actually contain, and rounding it to 595 x 842 moves the trim by 0.1 mm on one axis and 0.04 mm on the other. The imperial sizes are the same problem from the other end: US Letter is exactly 612 x 792 pt and untidy in millimetres.

## What a bleed costs, and what fitting does

| Bleed | A4 page | Fit to A4 | Trim that comes out | Lost each way |
| --- | --- | --- | --- | --- |
| 0 mm | 210 mm x 297 mm | **100%** | 210 mm x 297 mm | 0 mm / 0 mm |
| 1 mm | 212 mm x 299 mm | **99.06%** | 208.02 mm x 294.2 mm | 1.98 mm / 2.8 mm |
| 2 mm | 214 mm x 301 mm | **98.13%** | 206.07 mm x 291.45 mm | 3.93 mm / 5.55 mm |
| 3 mm | 216 mm x 303 mm | **97.22%** | 204.17 mm x 288.75 mm | 5.83 mm / 8.25 mm |
| 5 mm | 220 mm x 307 mm | **95.45%** | 200.45 mm x 283.5 mm | 9.55 mm / 13.5 mm |
| 10 mm | 230 mm x 317 mm | **91.3%** | 191.74 mm x 271.17 mm | 18.26 mm / 25.83 mm |

**The fit scale is governed by the tighter axis**, which for a portrait page is the width — so a 3 mm bleed on A4 costs 2.78% and the finished piece comes out 5.83 mm narrow. Nobody notices until two issues are side by side on a shelf.

## Creep, which grows with the job

| Extent | Sheets | On 0.08 mm | On 0.1 mm | On 0.15 mm | On 0.25 mm |
| --- | --- | --- | --- | --- | --- |
| 8 pages | 2 | 0.08 mm | 0.1 mm | 0.15 mm | 0.25 mm |
| 16 pages | 4 | 0.24 mm | 0.3 mm | 0.45 mm | 0.75 mm |
| 32 pages | 8 | 0.56 mm | 0.7 mm | 1.05 mm | 1.75 mm |
| 48 pages | 12 | 0.88 mm | 1.1 mm | 1.65 mm | 2.75 mm |
| 64 pages | 16 | 1.2 mm | 1.5 mm | 2.25 mm | **3.75 mm** |
| 96 pages | 24 | 1.84 mm | 2.3 mm | **3.45 mm** | **5.75 mm** |
| 128 pages | 32 | 2.48 mm | **3.1 mm** | **4.65 mm** | **7.75 mm** |

Creep is `(sheets − 1) × caliper` at the innermost spread, and the content has to move inward by up to half of it. **The bold figures are bigger than a 3 mm bleed**, which means the inner pages are being trimmed into rather than around before the guillotine has been allowed any tolerance at all. A 96-page job on 0.15 mm stock creeps 3.45 mm — 115% of the bleed.

## What each butting style gives

| Trim | Shared bleeds | Separate bleeds | What separate costs |
| --- | --- | --- | --- |
| A4 210 mm x 297 mm | **2-up** landscape | 1-up portrait | 1 of 2 |
| A5 148 mm x 210 mm | **4-up** portrait | 2-up landscape | 2 of 4 |
| A6 105 mm x 148 mm | **8-up** landscape | 4-up portrait | 4 of 8 |
| DL 99 mm x 210 mm | **6-up** portrait | 3-up landscape | 3 of 6 |
| 148 mm square 148 mm x 148 mm | **4-up** portrait | 2-up portrait | 2 of 4 |

On SRA3 320 mm x 450 mm less 10 mm of gripper edge and 5 mm of margin, which leaves 310 mm x 435 mm to print on. **Two pages that share a trim edge share one cut and one bleed**, so the pitch is the trim size. Two pages with their own bleeds need a gutter of at least twice the bleed, so the pitch is trim + 6 mm + the gutter — and on a sheet that is nearly full that is the difference between an N-up and half of one.

## What each sheet holds

| Sheet | Size | Printable | A5 shared | A5 separate |
| --- | --- | --- | --- | --- |
| A4 | 210 mm x 297 mm | 200 mm x 282 mm | 1-up | 1-up |
| US Legal | 215.9 mm x 355.6 mm | 205.9 mm x 340.6 mm | 1-up | 1-up |
| Tabloid | 279.4 mm x 431.8 mm | 269.4 mm x 416.8 mm | 2-up | 2-up |
| A3 | 297 mm x 420 mm | 287 mm x 405 mm | 2-up | 2-up |
| SRA3 | 320 mm x 450 mm | 310 mm x 435 mm | 4-up | 2-up |
| SRA2 | 450 mm x 640 mm | 440 mm x 625 mm | 8-up | 6-up |
| B2 | 500 mm x 707 mm | 490 mm x 692 mm | 9-up | 9-up |

**The N-up moves in whole steps**, so the useful question is never how much bigger a sheet is but which of the ones you can buy crosses the next step. And the printable area is always smaller than the sheet: 10 mm comes off the gripper edge once and 5 mm off each of the other three.

## Thresholds

| Constant | Value | What it decides |
| --- | --- | --- |
| points to the inch | 72 | a definition, and the reason nothing metric is round |
| mm to the inch | 25.4 | also a definition, so a point is 0.352778 mm exactly |
| a thin bleed | 3 mm | below this the bleed is reported as leaving the cutter too little |
| crop-mark offset | 3 mm | how far outside the trim a mark starts |
| crop-mark length | 5 mm | so the marked box is trim + 2 x (bleed + 3 mm + 5 mm) |
| poorly used | 60% of the printable area | below this the waste is reported |
| creep worth warning about | 50% of the bleed | above this the creep is reported as eating it |
| assumed bleed | 3 mm | used when JOB leaves it out, and always reported |
| assumed stock | 0.1 mm | used when a stitched job leaves it out, and always reported |
| assumed gripper edge | 10 mm | comes off the usable length ONCE |
| assumed sheet margin | 5 mm | comes off the width twice and the length once more |
| assumed butting | `shared` | the optimistic reading, and what a same-trim job normally does |

**The first two are definitions rather than choices**, and they are in this table so a reader can check them rather than take them on trust. The rest is this page's policy, and the harness proves every one of them is load-bearing by moving it and requiring the report to move with it.

## Sheet grammar

Two blocks, both `key | value`, one per line.

```
JOB
name      | the 64-page programme
trim      | a5          a name, or `148 x 210`, or `5.5 x 8.5 in`
bleed     | 3 mm        a bare length is MILLIMETRES; pt and in are converted
pages     | 64          the extent
copies    | 2500        the run length
stock     | 0.1 mm      the CALIPER, not the weight
binding   | saddle      saddle, perfect or none
marks     | yes         whether crop marks have to fit outside the bleed

SHEET
sheet     | sra3        the press sheet
grip      | 10 mm       the gripper edge, unprintable, off the length ONCE
margin    | 5 mm        the other three edges
butt      | shared      shared or separate
gutter    | 6 mm        only used when the bleeds are separate
```

**A BARE LENGTH IS MILLIMETRES**, because that is what a print job is spoken in. `pt`, `in`, `cm` and `"` are all accepted and converted — and a `pt` figure is converted rather than treated as a millimetre, because the two get mixed constantly and one of them is 0.35 mm.

**A TRIM IS A NAME OR TWO NUMBERS.** The names the page knows are `a3`, `a4`, `a5`, `a6`, `b2`, `sra3`, `sra2`, `letter`, `legal`, `tabloid`, `dl`, `square148`. Two numbers take an optional unit — `8.5 x 11 in` is read as inches and reported in millimetres and points.

**`stock` IS THE CALIPER AND NOT THE WEIGHT.** Two papers of the same grammage can differ by half again in thickness, and creep is sheets times caliper — so a thickness derived from a weight is a guess this page cannot improve on.

**`butt` DECIDES THE PITCH.** `shared` puts the trim edges together so one cut serves two pages and the bleed exists only around the outside of the block; `separate` gives every page its own bleed and forces a gutter of at least twice it. The gutter is ignored when the bleeds are shared, because there is nothing between the pages to put in it.

**THE GRIP COMES OFF ONCE AND THE MARGIN TWICE.** A press holds one edge, so the printable area is `(sheet width − 2 × margin) × (sheet length − grip − margin)`.

## Lanes

- **`plan`** — Specify the page from the trim, not the paper. A job gets specified as a paper size, and the file then arrives at the trim with no bleed, or at the media box and gets fitted back down. This works the other way: name the trim, decide the bleed against what the cutter and the binding will spend, and derive the page the file has to be — in points, because that is what the file is written in.
  Sections: Summary, The Sheet, The Setting, Reasoning, Next Step.
- **`check`** (primary) — Whether the file is the size it has to be, and what fits. The paid read of what the free panel computes. The media box against the trim, exactly what fitting would do to it, the creep the binding will spend out of the bleed, how many the sheet holds and what it wastes — and which of those is actually the problem.
  Sections: Summary, Verdict, Findings, Corrected Sheet, Next Step.
- **`boxes`** — The box question: what the page has to be, and in what unit. A point is exactly 1/72 inch, so no metric page is a round number of them — and the page is not the trim. This reads the geometry apart: the trim, the media box, the box the marks need, and precisely what `fit to page` does to each of them.
  Sections: Summary, The Boxes, What Fitting Does, What The Bleed Buys, Next Step.
- **`layout`** — The sheet question: what it holds and what it wastes. A press grips one edge and cannot print there, and two pages that share a trim edge share a cut. This reads the imposition: how many fit each way round, what butting them costs, what is left over, and what the next sheet size up would give.
  Sections: Summary, How Many Up, What The Sheet Wastes, What A Change Would Give, Next Step.
- **`deliver`** — Decide what changes: the file, the sheet, or nothing. Sorts every finding into what changing the file fixes, what only a different sheet fixes, and what nothing fixes. The third bucket is the honest one: a point is 1/72 inch, creep is the paper's thickness, and nobody downstream can be stopped from pressing `fit to page`.
  Sections: Summary, A File Change Fixes, Only A Different Sheet Fixes, Nothing Fixes, Next Step.

## Findings

All 33 are computed in the browser and cost nothing.

| Code | Severity | Scope | What it says |
| --- | --- | --- | --- |
| `NO-TRIM` | error | job | No trim size, so there is nothing to impose |
| `TRIM-NOT-POSITIVE` | error | job | A trim size with no width or no height |
| `NOTHING-FITS-THE-SHEET` | error | sheet | The page does not fit the press sheet even once |
| `FIT-TO-TRIM-WOULD-SHRINK-IT` | warn | job | Anything that fits this page to the trim size shrinks the artwork |
| `CREEP-EATS-THE-BLEED` | warn | job | The creep on this booklet is a large share of the bleed |
| `SEPARATE-BLEEDS-COST-YOU-UP` | warn | sheet | Butting the trim edges would fit more on the sheet |
| `MARKS-DO-NOT-FIT` | warn | sheet | There is no room on the sheet for the crop marks |
| `THE-SHEET-IS-POORLY-USED` | warn | sheet | Most of this sheet becomes waste |
| `THE-BLEED-IS-THIN` | warn | job | This bleed leaves the cutter very little to be wrong with |
| `THERE-IS-NO-BLEED` | warn | job | No bleed at all, so any cut error shows as white |
| `PAGES-ARE-NOT-A-MULTIPLE-OF-FOUR` | warn | job | A saddle-stitched extent has to be a multiple of four |
| `BLEED-ASSUMED` | warn | job | No bleed was given, so one was assumed |
| `STOCK-ASSUMED` | warn | job | No paper thickness was given, so one was assumed |
| `GRIP-ASSUMED` | warn | sheet | No gripper margin was given, so one was assumed |
| `MARGIN-ASSUMED` | warn | sheet | No sheet margin was given, so one was assumed |
| `BUTT-ASSUMED` | warn | sheet | How the pages are butted was assumed |
| `SHEET-ASSUMED` | warn | sheet | No press sheet was given, so nothing was imposed |
| `A-POINT-IS-A-DEFINITION` | note | job | A point is exactly 1/72 inch, which is why nothing metric is round |
| `THE-TRIM-IN-POINTS` | note | job | What the trim size is in the unit the file is written in |
| `THE-MEDIA-BOX` | note | job | The page the file has to be, which is not the trim |
| `THE-MARKED-BOX` | note | job | What the page has to be if crop marks go outside the bleed |
| `THE-FIT-SCALE` | note | job | Exactly what `fit to page` would do here |
| `SHEETS-AND-CREEP` | note | job | How many folded sheets, and how far the inner ones creep |
| `THE-SHINGLE` | note | job | How far the inner pages' content has to move to survive the trim |
| `HOW-MANY-UP` | note | sheet | How many pages the sheet holds, both ways round |
| `THE-BLOCK-ON-THE-SHEET` | note | sheet | How big the imposed block is, against what the sheet can print |
| `WHAT-IS-LEFT-OVER` | note | sheet | What the sheet has left, and what it is worth |
| `SHARED-VERSUS-SEPARATE` | note | sheet | What each butting style costs on this sheet |
| `PRESS-SHEETS-FOR-THE-RUN` | note | sheet | How many press sheets the whole run takes |
| `THE-GRIP-IS-NOT-PRINTABLE` | note | sheet | What the gripper edge and the margins cost before anything is imposed |
| `ROTATING-WOULD-HELP` | note | sheet | The other orientation fits more |
| `PERFECT-BINDING-DOES-NOT-CREEP` | note | job | Perfect binding has no creep and a spine instead |
| `A-BIGGER-SHEET-WOULD-GIVE` | note | sheet | What the next sheet size up would hold |

## What this page cannot do

This page computes from the numbers on the sheet. It has not opened a file, measured a sheet of paper or spoken to a press.

- **The gripper and margin figures are conventions, not specifications.** Every press has its own printable area, and a digital press's is often published to the tenth of a millimetre. Where a layout is close, get the real numbers rather than these.
- **The imposition is a flat grid**, which is what digital work and most cut-and-stack work is. A saddle-stitched booklet is imposed as PAIRED pages in a folding scheme — 2-up per side with the page order interleaved — and the sheet count here is the flat one. The creep figures are right; the layout figures describe a flat grid of the same page.
- **Creep is the standard `(sheets − 1) × caliper` model.** Real creep depends on how the paper folds, its grain, how hard it is pressed and whether the stitcher crushes the spine, and a bindery will have its own allowance. This is the figure to design against and not a measurement.
- **Caliper is not weight.** Two stocks of the same grammage can differ by half again in thickness depending on how they are made, so a caliper taken from a weight is a guess. Ask for the caliper.
- **Perfect binding is not modelled beyond having no creep.** It needs a spine width from the caliper and the extent, and a hinge allowance, and neither is computed here.
- **Crop marks are assumed to be 3 mm offset and 5 mm long.** Registration marks, colour bars, slug lines and job numbers all need room too and none of them is counted.
- **Nothing here knows about the artwork.** Whether the elements inside the trim survive a 3 mm cut depends on where they are, and a safe margin inside the trim is a design decision this page has no view on.
- Nothing here reaches the network, reads a file, or opens a document.

## API

`POST https://api.skillsafe.ai/v1/app-api/run` with a bearer token from
https://impose-desk.skillsafe.ai/tokens.html. The body IS the input object — there is no
`input` wrapper and no `X-App-Slug` header. `task` is required and must be one of plan, check, boxes, layout, deliver. `POST /estimate` is free and validates the same body.
Full documentation at https://impose-desk.skillsafe.ai/api.html.

## Provenance

Lanes derived from the `nano-pdf` skill in https://github.com/steipete/clawdis.
The arithmetic, thresholds, sheet grammar and findings are this app's own.
Not affiliated with or endorsed by the authors of that repository.