Each SVG has a title and description; each corresponding PDF is single-page A4 with selectable text and embedded DejaVu font. Use these complete descriptions for a read-aloud, reflowed text or locally made tactile chart. The standard printed PDF/SVG is not a raised tactile/Braille asset. Box order and labels carry meaning without colour. The original data are in Sources A–F; public check values are separate.
Complete register · SVG
Heading: “COMPLETE REGISTER”. Subtitle: “Source B · 48 invented paper mock-ups · fill every cell and margin.” A blank table has rows Hatch, Ring, Solid, All paper mock-ups; columns Pattern family, Pass, Redraw, Total. Footer: “Count in two directions / Pass + redraw = all. Row totals = all. / A pass means only the internal paper-border check.” Tactile: make four raised row mats, three labelled column slots, and count tiles; retain the Solid row.
Denominator windows · SVG
Heading: “DENOMINATOR WINDOWS”. Subtitle: “Move the category boundary; keep the fictional source cells fixed.” Top box FRAME 1 · ALL PAPER TILES says “Hatch + Ring + Solid / Passed _ of all _ = _ % / Included rows: Hatch, Ring, Solid / What was observed? One fictional desk print check.” Lower box FRAME 2 · SELECTED FAMILIES says “Hatch + Ring only; Solid omitted / Passed _ of selected _ = _ % / Which source cells changed? None. / Why can the second frame not stand for all tiles?” Footer: “A changed denominator is a changed question.” Tactile: place the same count tiles into all-row and Hatch/Ring-only frames; do not remove the Solid records from the source pile.
Method / evidence mat · SVG
Heading: “METHOD / EVIDENCE MAT”. Subtitle: “A clear protocol is a plan until someone actually follows it.” Four boxes: QUESTION + RULE: “What exactly counts as a border pass? Write the 5 mm rule before seeing results.” RECORD CONTROLS: “Print file, paper stock, setting, orientation. Name and label each mock-up.” INDEPENDENT CHECK: “Second reviewer uses same written rule. Keep disagreements visible, do not erase them.” OBSERVED OR PROPOSED?: “Old fictional register: supplied source only. New result from your plan: NOT OBSERVED.” Tactile: four numbered procedure strips; a token enters an observed-results box only after an actual authorised action.
Layout contrast · SVG
Heading: “TWO WAYS TO FRAME ONE SOURCE”. Subtitle: “Original fictional layout drafts for a visual hierarchy critique.” First upper panel DRAFT SPOTLIGHT · first glance shows large “64%”, then “READY FOR PUBLIC DISPLAY”; small caption “23 of 36 selected Hatch/Ring tiles passed a fictional paper-border check.” It notes missing Solid row and visitor evidence. Second lower panel DRAFT FULL REGISTER · first glance begins “25 of 48 paper mock-ups passed the single-print border check (≈52.1%).” It adds “Hatch/Ring subgroup: 23/36 (≈63.9%). Solid: 2/12. No visitor or accessibility test. Decision: consider another documented paper check.” Bottom question: “Which arrangement keeps the limit where readers first look?” The first draft is deliberately misleading as a teaching source, not a real promotion. Tactile/voice: read each panel's first line before its smaller text; compare what a reader can know at that stage. Do not claim an actual viewer reaction was tested.
Editor brief canvas · SVG
Heading: “EDITOR BRIEF CANVAS”. Subtitle: “Create an original visual report; this page is a scaffold, not a finished artwork.” Blank PANEL ONE · FULL SOURCE asks “Paper-border result: _ of _ / What exactly passed? Which table?” Blank PANEL TWO · DECISION + CHECK asks “Subgroup if needed: _ of _ / Method limit + next paper action + who does it.” Both have ruled space. Footer: “No panel may imply public approval or tested access.” Tactile/voice: compose two separately labelled panels or segments, evidence before proposal; Visual Arts evidence needs a created visual composition, not merely a spoken explanation.
The offline filter has the same paper denominator route. No generated aid turns fictional counts into a real result.