# Review record ยท Small world, big ideas ## Authoring scope Ten parent-supported play starts, 30 task-specific routes, 20 worked fictional contexts, ten later starts and ten extras. Every card has a fresh entry, including on days where earlier work is unavailable. A roughly ten-minute path can be split, changed, repeated or stopped. It is not a full early-years programme, age test, compulsory two-week sequence or evaluated intervention. ## Independent parent/teacher critique An author assigned to a separate primary-year band read all ten source starts, 30 routes, 20 examples and all five parent/educator guidance documents in full. They independently checked the current approved-frameworks and Outcome 4/5 primary HTML. No substantive framework/evidence blocker was identified; five concrete workflow/clarity findings were acted on: - Every card now has a one-sentence opening of twelve words or fewer. Longer invitations are optional adult prompts, rather than required child language. - Starts 05/07/08/09 begin with distinct playful actions: help a visitor, make a new pretend event, welcome a visitor, choose a making mode. - AAC is expanded in the standalone parent entry documents. - No-material means no additional play materials; usual communication supports remain available. A named-idea/established-signal/wait/confirm route makes nonverbal no-print selection actionable. Unclear is not agreement. - Applicable approved service frameworks and local requirements must be checked and followed; their adoption is not a discretionary pack choice. The quickest paper/oral start now precedes the optional helper link. The reviewer did not inspect or claim passes for unfinished prints/helper. This was an adult desk critique, not a family or learner trial. ## Actual authoring and artifact checks - Default checks passed ten complete starts, 30 concrete routes, 20 distinct examples, ten later starts and ten extras; every loose range clock totals ten minutes without gaps/overlaps. These are unmeasured optional paths. - All eight actual stored PDF pages were rendered and individually viewed in full. Broad spaces, supplied shapes, labels and exact linear equivalents agree; no clipping or filled child result was found. The PDFs remain untagged. - Genuine temporary generation reproduced 29 outputs byte for byte: ten cards, read-aloud scripts, eight SVG/PDF pairs, full alternatives and helper. All authored SHA/size/mtime values stayed unchanged during default checks. - Twenty-seven temporary meaningful faults were rejected at their intended gates after renewed item hashes: clocks, missing/thin/duplicated routes or examples, fictional feedback, untried benefit, erased adult attribution, completion rewards, source version/URL, outside collection, no-script controls, malformed/unbuilt workflow, page bounds, actually re-rendered changed print, lost linear equivalent, links/fragments and upstream notice. Three ordinary-cache positive controls passed. Faults are not learning tests. - The exact upstream DejaVu notice (3,859 bytes) is retained. All three Python files pass the repository's Ruff checks; 13 parenthesisation-only formatting repairs were independently shown to preserve the entire Python syntax tree. ## Actual interaction checks Isolated Chromium 151.0.7922.34 checks passed at 320, 390 and 1,024 pixels: 81 selected place/visitor/need combinations, 30 complete guide views, keyboard activation, touch choices, same-choice clearing, reset and blank reload; three no-script states retained disabled controls and all ten linked cards. Selected state and meaningful scene descriptions agree. Native buttons/select have at least 44-pixel targets; document/controls remain contained. No errors, outside requests, input fields, cookies or local/session storage were found. Print-media hiding was checked; no actual print dialog, browser-exported PDF or screen-reader trial is claimed. The owner also used the actual in-app Browser at 320 and 1,024-pixel overrides: keyboard/pointer scene choices, a different play start and expanded complete parent routes were visibly checked. That review exposed a long scroll before the first invitation; the current short opening now sits above the scene and changes with the selected play start. The unchanged artwork and full guides remain available. The updated helper passed the same isolated interaction checks including each of the ten top invitations. The desktop review also exposed an unnecessarily stretched picture-panel border alongside expanded guidance; panel alignment now keeps its natural height. Actual Browser logs had no warnings/errors at that check. Native screenshots are private review evidence; the capture backend may reflow to accommodate its scrollbar, so they are not claimed as exact viewport pixel-fidelity evidence. ## Four role review **Parent:** The one-sentence entry, no additional play materials, familiar communication route and no homework/record burden are practical provisions. Longer cards and example language stay optional; actual time, interest and use with an individual child still need a family trial. **Child:** Choices make a visible supplied scene immediately. One choice, an own paper/story idea, repeat or finish is valid. This is a small source of play ideas, not an adaptive tutor or proof that children find it fun. **Educator:** Intentions, actual evidence/fiction/adult distinctions and contextual next moves are explicit. No graded developmental rubric, school code award, whole-outcome certification or local-service adoption is inferred. **Funder:** Original editable source, rights, item hashes, actual artifact and interaction checks are reviewable. These provide execution evidence only; uptake, classroom effectiveness, inclusion outcomes and savings need genuine human evaluation. No numerical quality or viral-potential claim is made. ## Human review remains open No parent, educator, child, community language speaker, assistive-technology user or funder has trialled this pack. Desk and Browser checks cannot show engagement, pacing, developmental benefit or individual access. Source links do not prove local framework adoption or full outcome coverage. Candidate future narration would need actual listening and transcript review before any reviewed-audio claim.