Packaging artwork review. One label proof on a canvas, annotated, with its colour separations. Brief #10.
Direction
| Lever | Decision |
|---|---|
| Archetype | Canvas. The tool rail is the only permanent chrome. The dock ships open, since the thread is the content, and closing it hands its width back to the proof: 40 % → 89 % at 375 px, 95 % → 102 % at 1280. Above 1400 px the fit is height-bound and closing it changes nothing but the pasteboard, which is what a portrait document does. |
| Data stance | Prose-first. An entry leads with the sentence somebody wrote and carries the replies to it; note 1 is an argument between prepress, account and label compliance, not a finding. The selection is shared: picking an entry draws its measurement on the artwork, picking a pin brings its entry up. |
| Density | Spectrum desktop scale: 32 px controls, 14 px body. The Interface scale control switches the whole page to the mobile scale (40 px, 17 px). |
| The one bold thing | Keyboard access designed rather than inherited. Arrows pan, +/- zoom, 0 fits. Escape closes the dock and returns focus to whichever control opened it. |
Second state
Desktop → mobile is a whole-interface rescale, not a touch-target patch. spectrum_tokens.py --platform-diff reports 922 of 922 platform tokens differ. The page carries the ones it uses:
| Token | Desktop | Mobile |
|---|---|---|
component-height-100 | 32px | 40px |
component-height-75 | 24px | 30px |
font-size-100 | 14px | 17px |
font-size-75 | 12px | 15px |
line-height-font-size-100 | 18px | 22px |
workflow-icon-size-100 | 20px | 24px |
text-to-control-100 | 10px | 13px |
text-to-visual-100 | 6px | 8px |
Measured in Chromium at all three widths: the Undo control goes 32 → 40 px and the body font 14 → 17 px. The page reads the scale from (pointer: coarse) and the control overrides it, which is what the skill asks for.
The sheet is a separate decision, and it is made on width. It was made on the scale, and at 1440 px the Mobile control therefore produced a 1383 px-wide sheet lying over the canvas with the zoom stack underneath it, which is the one thing SKILL.md line 52 forbids ("nothing permanently occludes the work"). Below 760 px the dock floats; the canvas counts the height it takes as spent, so the artwork centres above the sheet and the zoom stack rides on top of it. At 1440 the mobile scale is a rescale and the dock stays a 364 px dock.
Generator
spectrum_tokens.py --both > tokens.css
spectrum_tokens.py --platform-diff
spectrum_tokens.py --theme desktop --format css
spectrum_tokens.py --theme mobile --format css
spectrum_tokens.py --theme default --format css
spectrum_tokens.py --contrast
The generator wrote tokens.css; nothing here edits it. --both emits the colour axis only (671 tokens per set), so the page declares the sizing tokens it uses in its own <style>, from the desktop, mobile and default runs.
Zoom
The artwork is 90 × 130 mm, drawn at 340 px per 90 mm (96 dpi), so --zoom: 1 is the label at print size. On load the page fits the artwork to whatever canvas the open dock leaves, measured: 40 % at 375, 44 % at 390, 95 % at 1280, 119 % at 1440, 149 % at 1920. + and - step the 25 / 50 / 75 / 100 / 150 / 200 / 300 / 400 / 600 ladder from there and disable at its ends, 0 refits, and a resize (or opening, closing or resizing the dock) refits only until the reader picks a zoom. The pan is clamped so 48 px of the proof is always on screen: a canvas that can lose the document has lost it, and nothing on the status bar would say so.
Guides
The pasteboard carries trim marks and the 3 mm safety margin, and the selected pin draws its own measurement on the proof. That is what closes the gap between the thread and the canvas: note 1 says the net weight sits 1.4 mm from the trim, and the artwork now visibly breaks the safety margin at that corner. Without the guides "Show pin 1" moved a coloured dot and the reader had to take the sentence on trust.
The box, the leader and the dimension ticks are in the artwork's coordinate system and scale with it. The label is too (font-size: 3px, three millimetres of proof), but it carries a floor. a11y-probe.mjs --width 390 measured the first version at 7.7 px: 3 units times the 44 % fit is illegible, and a measurement nobody can read is not a measurement. --anno-k counter-scales the glyphs through the scale property, so the label grows with the proof and stops shrinking with it at --anno-floor: 12 px desktop, 15 px mobile, alongside every other sizing token. Measured: 12 px at 390 and at 1280, 13.5 px at 1440, 16.9 px at 1920, 34 px at 300 %. The floor is on the type only; the geometry still belongs to the page, which is what Acrobat and every other proofing tool do with review markup.
Measured contrast
| Pair | Dark | Light |
|---|---|---|
neutral-content-color-default on background-layer-1-color (dock prose, 16px) | 12.44:1 | 13.70:1 |
accent-content-color-default on background-layer-1-color (links, 14px) | 4.91:1 | 4.53:1 |
--on-solid on accent-background-color-default (primary button, group pin, selected pin): gray-1000 in dark, gray-25 in light | 4.51:1 | 4.81:1 |
gray-600 control boundary on layer 1 | 4.99:1 | 4.60:1 |
negative / notice / positive visual marks | 4.90–4.92:1 | 3.72–3.73:1 |
Every pair clears WCAG 2.2 AA. An earlier draft of this table claimed gray-25 on the accent in dark at 4.51:1 and the light control boundary at 3.31:1; both were wrong when computed from tokens.css (4.19:1 and 4.60:1), so the label now takes --on-solid, which flips per theme. The three accent pairs fail APCA at 14 px; that is Spectrum's own semantic accent, which the system's own audit records at 4.81:1 with the note that it passes "without much room". No accent-coloured text on this page is smaller than 14 px.
Deviations from CONTRACT.md
- Raw hex inside the artwork and the ink swatches. The label SVG and the six plate chips carry the customer's colour data. Spectrum's own rule for a creative tool forbids a UI theme reaching into it (
references/color-theming.md§ Never color-manage the user's content). Every pixel of chrome is avar(--spectrum-*). - Sizing tokens declared in the page rather than linked.
--bothemits no sizing token; the values are the verbatim output of the three runs in the Generator section. - Durations authored at 90–190 ms.
spectrum_tokens.py --grep durationreturns nothing; the set carries no motion token. The range is the system's published 100–250 ms (references/scales-typography.md§ Motion). - Source Sans 3 in place of Adobe Clean, which Adobe licenses to its own products. The stack names Adobe Clean first.
- Two
copy_lintrules kept, recorded incopy-lint.ignorewith the reason per occurrence:shouty-capsandtitle-casefire on the printed cider label inside the SVG, on three one-word table headers read as one string, and on a heading that is already sentence case.
Uplift, 2026-10-29
What changed from the first version, and why each part is a decision rather than decoration.
- The Annotate tool is gone. Comments are not placed by dragging; the reader chooses a pin or the Properties panel and changes the thing the pin measures.
- The one interaction is the proof itself. Clearance, type size and the black build are live properties. A handle on the artwork, a slider, a number field and the arrow keys all write the same value; the pin status, the measured sentence, the Approve button, the Separations table and the heaviest-ink meter all follow, and a live region announces the result. Every drag has a keyboard and a click equivalent (WCAG 2.5.7).
- History is a named, coalescing snapshot stack (Undo and Redo with the next step's name, Ctrl+Z, Ctrl+Shift+Z) and is listed in Properties with times; any row is a jump.
- Rulers and a measure tool are drawn in millimetres and snap to device pixels; the crosshair and its two ends are keyboard-operable.
- Plate figures are derived, not typed. The Separations panel rasterises the artwork SVG and measures each plate's share of the 90 x 130 mm trim area (the barcode is excluded in a second pass because it prints from black only). The figures depend on the fallback fonts the machine supplies for Georgia and Helvetica, because the rasteriser draws the SVG as an image. The EAN-13 is valid: parity LGGLLG for a leading 5, 0.33 mm module, check digit 9.
- Approve uses
aria-disabled, notdisabled, so the reason ("2 blocking comments to clear first") can be reached by keyboard and hover. - Generator command unchanged.
tokens.cssis not edited; the new--on-solidand--on-noticeare page-level aliases of Spectrum gray steps.
Round two, 2026-10-29 (first fold)
- The page opens at a working zoom, not a fit. The first view is the blocking comment's corner of the label at about 335 % (220 % at 390 px), with rulers, trim, bleed and safe area in frame. The artwork now runs 3 mm into the bleed (the viewBox is still the 90 x 130 trim box, so the plate measurement is unchanged). Fit and the 0 key still show the whole label. The canvas grid track is now the canvas itself, which is what centres an artwork larger than the view.
- A navigator (the label in miniature, four pin dots, the visible rectangle; click or drag to move) and a key to the three guides replace the empty pasteboard. The zoom ladder gained 400 % and 600 %.
- The one measurement is drawn on the art: a hatched 3 mm keep-out at the trim, the measured clearance as a dimension line, the shortfall against the 3.0 mm spec in the negative role, and a two-line label. All of it follows the slider, the handle and the number field.
- The critic's "500 ml e" was the EU estimated-quantity mark set 1.4 mm from the trim, with the drag handle sitting on it. It is now the real mark (U+212E), and the trim, bleed and keep-out make the proximity read as the finding rather than a clip.
- The comments panel is a Spectrum accordion: one section per pin (pin, title, state, live value), field labels for Measured, Spec and Raised by, the sentence under them, replies folded. Opening a section selects its pin and brings it into view; picking a pin opens its section.
- The callout type is 2.4 mm (was 3 mm) with the same 12 px / 15 px floor.
Round four
- The page opens fitted, not at 335 %. A stranger now sees the whole Aster Hill cider label with pin 1 and its measurement callout in frame (111 % at 1440 by 900, 93 % at 390); the round-two working zoom is gone because the callout floor keeps the finding legible at fit. The fit leaves 36px clear of the guide key so the label's corner no longer sits under it.
focusViewand its constants are deleted. - The phone sheet opens as a peek. At 760px and below the panel starts 156px tall (grip, tabs, the first comment's title and state); the grip, any tab or any opener raises it to the full sheet and the grip lowers it again. The label keeps about two thirds of the screen at 390.
- The design notes sit on a twelve-column grid from 1101px: title and sentence beside the four decisions, the two paragraphs side by side, the three measurements across, and the changes in two columns, at the tool's own density. Below that width the old stack is unchanged.
Round eight
- Gap beside "The second state" and "Where the colour comes from" (1440, both themes). Real. The two sections sat in the two halves of the twelve-column grid, but each paragraph stops at 62 characters, so the first ended well short of the second's start and the second left a third of its half empty. The second state now sits under the title on the title's edges (columns 1 to 4); the colour section sits under the four decisions (5 to 12) in the same two columns they use, with the paragraph in the first and the four status marks stacked in the second. No column starts where another one's text ends. The light and dark themes use the same rules.
- Minimap over the label's top-right corner (390 light). Real. On a phone the fitted label fills the canvas, and the miniature sat on its corner to show a view rectangle that was the whole label. It is now hidden while the canvas shows the whole label and appears as soon as the label is zoomed past the canvas, when the rectangle means something (checked: hidden fitted, shown at 150%, and still hidden on the wide layout's fit view).
- Declined: nothing.
Round nine
Found by driving the phone as a coarse-pointer device (the review renders use a fine pointer, so they never showed it) and by raising the sheet.
- The callout broke on a phone. Its two lines were two separate text elements, each counter-scaled about its own centre, so at the mobile floor (15 px) they overprinted each other and the type. It is now one text block with two lines, so the leading scales with the type.
- A label that outgrows the art leaves it. Past a counter-scale of 1.7 (a phone fit, a zoomed-out view) the callout hangs below the trim marks, in the interface's own ink on the pasteboard, and its second line shortens to "short 1.6 mm". The panel carries the spec.
- The zoom stack no longer sits on the label. On a phone the fit reserves the strip the stack occupies, so the foot of the label (the web address, the ABV) is not hidden behind it. The stack sits 8 px above the sheet instead of 16.
- A raised sheet shows the comment, not a thumbnail. With the sheet up the canvas is a strip and the fitted label fell to 41 %, smaller than its own callout. When the fit is under 70 % and the sheet is raised, the view goes to the selected comment's marks at print size (100 %); picking another comment re-aims it, lowering the sheet or pressing 0 shows the whole label again. Zoom, pan or the navigator hand control back.
- The peek is exactly one comment. It was a fixed 156 px that cut through the next row; its height is now measured from the grip, the tabs and the first comment's title, at either scale.
- The app bar on a phone is two lines, not three. The job name had wrapped over two lines with its version alone on a third.
- The cursor readout shows no coordinates while the pointer is on the zoom stack, the navigator or the key.
- The net-contents box has 0.3 mm more room on its right so the estimated mark is not struck through.
- Declined: the pin sitting on the first letter of the ingredient line (it marks the line it is about; severity 1).
Round ten
A second, cold read. Round nine's 33/40 did not hold: the page's one commit, Approve, was a single press with nothing to confirm it, no way to reopen it and a reason for refusal that lived in a hover tip.
- Approve asks, and names what is still open (severity 3). Pressing it opens a Spectrum alert dialog: "Approve v7 for plates?", the open comment and the small-type finding listed (comment 4 with a "Show comment 4" link), "Keep reviewing" focused first, Escape and the backdrop cancel and focus returns to the button.
- A refused Approve says why, on a phone too (severity 2). With blocking comments the same dialog lists each, with its figure and a link that closes the dialog, opens the sheet to the comment and focuses it. The hover tip stays for pointers.
- A way back (severity 3). Once approved, the button becomes "Reopen proof" (outlined, with a return arrow instead of a tick, so a status mark no longer looks like a control). Undo still works, and the history lists "Reopen proof".
- The page changes after the act (severity 2). The app bar now carries the proof's status ("In review" / a tick and "Approved") beside the version; the status bar reads "In review · Saved 14:32", "Approved 14:35" or "Approval withdrawn · Saved 14:36". Editing an approved proof withdraws the approval and now says so in the live region and the status bar, where it used to drop the stamp silently.
- Comment 4 can resolve (severity 2). At 0 mm of varnish over the window there is nothing for the printer to confirm, so the comment reads Resolved and "none over", and the counts and the dialog follow. It used to stay Open at "0.0 mm over".
- The ink limit is explained beside the figure (severity 2). "Heaviest ink, all plates added: 264% of 280%" now says what 280% is, in the note under the meter.
- The finding follows the type size (severity 2). Above 7 pt the "four plates print the small type" claim contradicted its own rule; it now reads "The ingredient type is 7 pt or over", with the fix button relabelled.
- Phone details found while driving: the app bar job line wrapped to three lines once a status joined it (version and status are now no-wrap), and the status check was 4 px taller than the line it sat in.
Inventions, none contradicted by the brief: Reopen means only that the approval is withdrawn and the proof is back in review (every property was already editable); comment 4 resolves at 0 mm of overlap; the dialog's wording ("This marks v7 approved for plates") restates the existing stamp; the definition of the 280% limit is the general meaning of total ink and is not tied to a stock. No prices, no new clinical or legal claim.
Scores before and after, ten heuristics: visibility of status 3 to 4, match with the real world 3 to 3, control and freedom 3 to 4, consistency 3 to 3, error prevention 2 to 3, recognition over recall 3 to 3, flexibility 4 to 4, minimal design 3 to 3 (three bands under the sheet on a phone), error recovery 2 to 3, help 2 to 3. Total 28 to 33.
Left: the callout pin 2 still covers the first letter of "Dabinett" at phone fit (severity 1, declined in round nine); the disabled-looking Approve is an icon alone at 390 and relies on the dialog for its name (severity 1).
Round eleven (2026-11-10)
Three severity 2 defects from the independent score (31/40). Each was reproduced first (1440 light, 390 dark and 360 light as touch).
- A resolved comment still drew the red hatch. At 3.2 mm the label read "Meets the 3.0 mm spec" over a red-hatched keep-out. The measurement region now carries
data-metwhile the clearance meets the spec: the hatch gives way to a flat positive tint (20 %) and the dashed minimum line turns positive. Short of the spec nothing changes. The geometry and the callout are untouched. - Phone touch targets under the mobile control height. Pins and the drag handle were 30 x 30 and the stepper 36 x 36 on the mobile scale. A page-level
--hit-targetiscomponent-height-75on desktop andcomponent-height-100(40) on the mobile scale. Pins and the handle are that box; the pin's visible disc moved to a::beforeatcomponent-height-75, so the drawing on the label is the same size and a finger gets 40. The number field's 2 px border is now drawn over the field (::after), so each stepper is the full control height (40 on mobile, 32 on desktop) in a field that is no taller. - The "blocking comments to clear first" tip stacked under the dialog and covered the dock. It now shows on hover only for pointers that can hover, and on keyboard focus (
:focus-visible); a tap never shows it and it is hidden whenever a dialog is open. From 601px it sits beside the Approve button instead of under it, so it no longer covers the dock tabs or Close. It still covers part of the interface-scale control while it is shown.
Inventions: none. The 20 % tint and the --hit-target name are the page's own; the desktop stepper is 4 px wider than before (32 against 28).
Round thirteen (2026-11-10)
Reservation from the independent re-score (34/40): "phone pins hide the artwork text they annotate". Reproduced at 390 dark and 360 light (touch, mobile scale): pin 2 covered "Cider apples", "Dabinett" and the INGREDIENTS heading, pin 3 covered the barcode digit "5" (and the first bars at 360), pin 1 covered the estimated mark after "500 ml". The mobile-scale disc is 30 px, which at a 72 % or 51 % fit is 11 to 15 mm of label.
- Pins stand off in the margin on a leader. On the mobile scale each pin keeps its exact anchor (the same
left/topas before) but its disc moves sideways out of the artwork, pins 2 and 3 into the left margin and pins 1 and 4 into the right, joined to the anchor by a 2 px line (a 4 px light line under it, as the measurement marks do, so it reads on paper and pasteboard in both themes) that ends in a dot on the point. Nothing in the artwork is covered at 390 or 360 (measured: no disc rectangle intersects any text or bar rectangle). On the desktop scale the lead is 0 and nothing changes (24 px discs, measured clear of the type at 1440 and 1920). - Mechanism.
--pin-lead(0 on the desktop scale, 28 px in both mobile-scale blocks) and--side(-1 or 1, inline on each pin) drive the pin'stranslate, itstransform-origin(so hover scale and the arrival animation pivot on the anchor and the line never leaves the point) and the::afterleader.--ktrims one pin's lead: pin 1 is 1.1x so its disc clears pin 4 at 360 (they overlapped before), pin 4 is 0.8x. - Targets and linking unchanged. The button is still the 40 px box on the mobile scale (24 px minimum on desktop, measured); the leader is
pointer-events: noneand no target. Selecting a pin, or a note, still rings the pin and opens its note; the selected pin keeps its ring in the margin, and pin 3 moved by Undo carries its leader with it. The pan's 28 px keep-in margin is respected, so tapping a pin does not pan the label.
Inventions: none. The 28 px lead and the per-pin factors are layout, not product facts.
Follow-ups on the same reservation (coordinator, same round):
- Callouts no longer cover label text on a phone. Reproduced: at 390 and 360 the callouts for comments 2, 3 and 4 (and, raised sheet, 1 and 4) overprinted "Dabinett", "NUTRITION", "5.8% vol" and "500 ml". When the callout is counter-scaled past 1.7x (the phone fit) every comment's label now takes the one slot callout 1 already used, below the trim marks on the pasteboard, right-aligned (
.callouton the three groups,text-anchor: endin the tight rule). Only the selected comment's is drawn, so the slot is never shared. The bare "5.2 pt" reads "Type 5.2 pt" there, since it is no longer beside the type (r2Val, painted bypaintSpec). With the raised sheet (zoom 1, not tight) callout 1 uses its short spec line past 1.5x so it clears the web address, and callout 4 sits 1.6 mm higher, clear of "500 ml". Measured with a text-rectangle intersection for all four comments, light and dark, 390 and 360, peek and raised sheet: no callout intersects any label text or bar. - Touch hit areas are 44 px. On a coarse pointer
--hit-targetis 44 px whichever scale is chosen (:root:root:rootin apointer: coarseblock, so it outranks the scale blocks). Pins and drag handles measure 44 x 44 and the drawn disc stays 30 px (24 px on the desktop scale with a coarse pointer). Pin 1 has the plain lead and pin 4 a 0.7 lead (--k) so every 44 px box stays within the 28 px keep-in margin of the pan and the discs of pins 1 and 4 still clear each other at 360; tapping each of the four pins at 390 and 360 does not move the label. The 44 px boxes of pins 1 and 4 overlap each other by a few pixels (pin 1, later in the order, wins the overlap), as the 40 px boxes did. - Left alone. The other mobile-scale controls (tools, stepper, icon buttons) stay at the Spectrum mobile control height of 40 px. With the sheet raised at 360 the zoom stack can still sit over the first word of comment 4's callout (not label text); the raised view is the existing round-nine behaviour.
Inventions: the wording "Type 5.2 pt" (the panel names the comment "Ingredient type size").
Round fifteen (2026-11-10)
Independent score of round fourteen (34/40): one severity 2 and two reservations.
- Enter in a number field dropped focus (severity 2). Reproduced at 1440 light: Adjust on comment 1, Ctrl+A, 3.2, Enter,
document.activeElementwas BODY (the Enter handler calledblur()). Enter now commits through the same function the blur uses and keeps focus in the field with the value selected, for all three fields (clearance, type size, overlap; each driven with Enter, the value committed, focus on the field). The same fault in a neighbour: a stepper, Reset, Undo, Redo or a zoom button that disables itself while it holds focus (a stepper reaching its minimum, Reset after use) dropped focus to BODY;setOff()now hands focus to the number field, or to the other half of the pair, before it disables. Reproduced for the size stepper at 4.0 pt. - Pins over the artwork text at 360. Not reproduced on the touch / mobile scale: with each comment selected, at 360, 375 and 390 wide, 740 and 780 high, light and dark, sheet at its peek and raised, no pin disc, leader or callout meets a text or bar rectangle. It did reproduce on the desktop scale at the same widths (a 360 px window on a mouse, or the Desktop scale chosen on a phone): pin 2 over INGREDIENTS, Cider apples and Dabinett, pin 3 over the barcode digit, pin 1 over "500 ml", and callout 1 over the web address with the sheet raised. Below 76 % zoom the pins now stand off on their leader (20 px) whichever scale is set (
data-smallon the artboard, set besidedata-tight), and callout 1 takes its short line from 1.25x counter-scale instead of 1.5x. A second overlap found while sweeping values, on the touch scale at 360, 375 and 390: callout 2 ("6.0 pt") sat under its box and covered the NUTRITION heading from 6 pt up, which is the value a reviewer sets to resolve the comment. Counter-scaled past 1.25x it now stands beside the box, where the line is empty (placeR2); at 1440 it stays under the box. Swept type size 4 to 8, overlap 0.4 to 2 and clearance 0.5 to 5: no callout meets label text at 360, 375, 390 (and 1024, 1440). Measured clear for all four comments, three widths, two heights, both themes, both sheet positions, on the touch, the desktop scale and a fine pointer. The type-size drag handle (a 10 px square on the measured box edge, only for comment 2) still sits on the second line of the ingredient type by design: it is the control for that measurement, not a pin. - Text links 30 px at 360. Show pin, Adjust, the reply toggles, Reset and the Portfolio link are 44 px high and wide on a coarse pointer (
min-heightandmin-width44, centred, same type size). Measured at 360 and 390 over every tab and each open comment: none under 44.
Inventions: none. The 76 % threshold, the 1.25 threshold and setOff's choice of where focus goes are layout and keyboard behaviour, not product facts.
Round nineteen (2026-11-10)
Independent score of round eighteen (37/40): one severity 2 and three severity 1 items. The severity 2 reproduced at 1440 light before any edit; the reason at rest and the 0 key were read from the code (opacity 0 at rest; fit() with no say). Letters below are the reviewer's.
- Enter or a click on "Print small type from black alone" dropped focus to BODY (severity 2). The fix hides itself once it has run.
renderSepsnow moves focus to the finding's own heading (#finding-h,tabindex="-1", which now reads "Black alone prints the small type") before it hides the button; the existing live line ("Set text black to black only. Heaviest ink 124 percent of 280. 2 blocking comments left.") still announces the change. Measured: keyboard and click both leave focus onH3#finding-h; Ctrl+Z restores the button. - The same fault in two neighbours, found by sweeping every control that removes or hides itself. (1) History rows are rebuilt on every jump, so Enter on a row dropped focus to BODY: the row that held focus now hands it to the new current row (
renderHistory). (2) The comment handles hide when the selection clears, so Escape on a focused handle with the dock closed dropped focus to BODY:select()now sends focus to the proof canvas for any focused element it has just hidden, and says "Selection cleared.". Steppers, Reset, Undo, Redo and the zoom buttons were already covered bysetOff(round fifteen); the approval dialog's buttons already return focus. Driven at 1440, 1920, 390, 360, 768 and 620: no case ends on BODY. - Sev 1 (a), the reason beside Approve showed only on hover or focus. From 601px the reason ("2 blocking comments to clear first") is now a plain line of the app bar beside the button at rest, in the subdued content colour, no tip chrome; it is gone once nothing blocks, as the tip was. At 600px and below there is no room, so the tip keeps its hover and focus behaviour and a tap still names the reason in the dialog.
- Sev 1 (b), the 0 key refit without an announcement.
0now says "Artwork fitted to the view, N percent.", the same words as the Fit button (onefitAndSayfor both). - Sev 1 (d), the zoom ladder. The page's ladder is 25 / 50 / 75 / 100 / 150 / 200 / 300 / 400 / 600 and 600 is right for a 90 x 130 mm label drawn at 96 dpi (inspecting a 0.3 mm register mark), so the Zoom section above now names the real ladder.
Not done: the 28 px dock grip at 390 (reviewer's sev 1 (c), not in this round's list).
Inventions: none. The two announcement strings reuse the page's wording; "Selection cleared." and "Pin N selected." are new status words, not product facts.
Round twenty-one (2026-11-10)
- Sev 2, the app bar changed height with its words. Reproduced at 1440 light: 49 px on load, 87 px after Adjust on comment 1 and one + press on the clearance stepper (Undo's label went from "Undo move pin 3", 153 px, to "Undo set clearance to 1.5 mm", 228 px, and the tool group wrapped), which moved the canvas and the dock 38 px. At 1280 and 1024 the bar was already two rows (87 px) from load while the reason beside Approve showed, and 99 px on the Mobile scale. Fix: (1) Undo shows the one word "Undo" at every width and the action it would undo moves to the tooltip below the button (the rail's
.tippattern: hover, keyboard focus, Escape hides, the next pointer move or focus brings it back) and to the accessible name (#undo-labelis now ansr-onlyspan, the visible word isaria-hidden); Redo got the same tooltip. (2) From 1024 px the bar is one row that does not wrap: the job name takes an ellipsis (full name intitle), the controls do not shrink. (3) Below 1440 the reason beside Approve takes two short lines (22ch) instead of one long one; from 1024 to 1199 the job, version and status stack as two 12 px lines (exactly one control high); from 1024 to 1099 the theme button is its icon alone, as on a phone, its name unchanged. Measured at 1024, 1100, 1200, 1280, 1366, 1440 and 1920, light and dark, on load and after Adjust, a stepper press, clearance to 3.0 mm, type size 6.0 pt, the approve dialog, approve, reopen, Ctrl+Z and the theme switch: the bar is 49 px and the canvas top is 49 px in every state. On the Mobile interface scale it is 57 px in every state (a scale is the reader's choice; its controls are taller). On the Mobile scale from 1024 to 1440 the job name is cut by an ellipsis because the type is larger; that is the cost of a bar that never reflows. - Sev 1, the 390 dock grip was 28 px tall. The grip is now 44 px tall with the same 36 x 4 handle drawn in it, so the sheet's peek is 16 px taller (the peek height is measured from the grip, so the first comment still fits whole) and the label has 16 px less canvas on a phone.
- Sev 1, the BRIEF's zoom ladder. Not reproduced: the Zoom section (round nineteen) already names 25 / 50 / 75 / 100 / 150 / 200 / 300 / 400 / 600, which is the page's
STEPS. The "34 px at 300 %" line in the guides section is a measured annotation size at that zoom (3 mm of label type at 300 % is 34 px), not a statement about the top of the ladder, and stays. Rounds two and nine name the ladder as it was then and are history. Round nineteen's "From 601px the reason is a plain line" is still true except that below 1440 it is two lines.
Inventions: none. The tooltip shows the history label the page already writes plus the two shortcuts the page already binds (Ctrl Z, Ctrl Shift Z).