# Design critique and refinement record ## Visual-interface copy revision Updated the first approach’s visual screenshot to “Your ideas, / with a visual interface” in both typography variants. Re-exported the screenshots and updated the copy deck and comparison boards. ## Larger, lower product compositions Increased the product scenes by 8% and moved their top edge 24 px lower. This uses more of the lower canvas while retaining the glass frames, centered headlines, both font versions, and the opening footer logo. Gallery image URLs now include a revision so refreshed pages load the latest exports. ## Website glass-frame revision Reused the website’s optical frame shader, current glass material settings, rim, and camera-island styling. Replaced the metal phone overlays and applied the glass treatment to dedicated-app cards. The still compositions use fixed lighting on white. Checked all 70 previews for successful optical rendering, missing images, metal-frame removal, and preservation of the typography, spacing, and footer-branding rules. ## Layout and branding revision Moved the headlines upward and placed the first visible artwork 38 px below each headline block at the 430 px design scale. Removed the repeated top wordmarks. Only the opening screenshot in each approach has a small footer wordmark. Product UI branding remains part of the reconstructed scenes. All 70 compositions were checked for the new gap, centered regular-weight headlines, absence of subheadlines and periods, and opener-only footer branding. Both complete series contact sheets were visually reviewed. ## Headline punctuation revision Removed periods from all 35 main headlines in both typography versions, while keeping the copy, line breaks, alignment, and typeface rules. Refreshed the 70 PNGs, 70 SVGs, review boards, gallery, and headline copy deck. ## Typography revision Updated all five approaches with two versions each: all Kalice, and Kalice on screenshot 1 followed by Selecta Regular on screenshots 2–7. Every headline is centered and all supporting subheadlines are removed. All 70 rendered compositions were checked for typeface, weight 400, centered alignment, missing assets, headline bounds, and absence of subheadlines. Both full contact sheets were visually reviewed. The 70 PNG exports were compared to the current editable source renders; all average pixel differences were under 5/255 (rendering antialiasing). The new gallery toggle switches comparison cards, complete series, and enlarged image links together. The notes below describe the original concept development before this typography revision. Three critique passes were run on rendered artwork, followed by export and gallery checks. These were self-review passes, not user testing or reviews by an independent panel. ## Pass 1 — Does each direction have a reason to exist? **Evidence:** all 35 screens were rendered at 430 × 932 and compared in five seven-screen contact sheets. **Findings:** the first layouts were too similar across directions. Several longer poetic headlines wrapped into the scene area. Some supporting labels repeated what the main copy already said. **Changes:** - The flagship opener now uses the current brand’s “Your personal agentic OS” positioning. - Small miracles received a dedicated action board with concrete completed steps. - A world that grows became more object-led; its visual and agent screens no longer depend on a phone mockup. - Our little world leads with two people and two views of their shared Movie Matcher. - Say it. See it. has a clear request, named result, and resulting app. - Long headlines were rewritten, and redundant bottom captions were removed. ## Pass 2 — Does the visual prove the promise? **Evidence:** a second complete render of all 35 screens, plus full-size inspection of the five opening screenshots. Font and headline/scene positions were read back from the rendered pages. **Findings:** translucent cards let unrelated interface text show through. The initial iOS scene mixed Lock Screen cues with a Home Screen arrangement. “Grows around you” needed a more concrete example. **Changes:** - Text-bearing cards now have solid, gently shaded surfaces; glass stays in the objects and device treatment. - The iOS screen uses a Home Screen grid with two small widgets, a wide flight widget, a second widget row, app icons, and a dock. - The growth example now asks to add a butterfly collection; the social version adds a shared Sunday soundtrack. - The flagship and object-led directions use a lighter silver device treatment. - The connection scene uses restrained colored category symbols and a clear 100+ count. ## Pass 3 — Does it hold up when inspected closely? **Evidence:** refreshed contact sheets and inspection of the full-resolution output, including the SVG versions. **Findings:** a few floating widgets crossed an app title; some result cards obscured part of a movie name. The first full-page export method produced incorrect internal scaling despite reporting the correct file dimensions. **Changes:** - Flight cards were repositioned so “Copenhagen, at our pace” stays intact. - Unnecessary result/proof cards were removed where they crossed the movie title. - Movie Matcher uses the original popcorn icon from the brand assets. - The export method was replaced with native-resolution sectional rendering. The final PNGs were assembled at 1290 × 2796 without upscaling a thumbnail. - SVG headline and supporting-copy baselines were measured from the rendered HTML. Their text remains separate from the product artwork. ## Final checks - 5 approaches × 7 themes = 35 screenshots. - All PNGs are RGB and exactly 1290 × 2796. - All final headlines fit above the artwork; font families match the intended Kalice/Selecta assignment. - No missing image or font loads were reported in the final render audit. - All five complete series and the opening comparison were visually inspected. - SVG text and artwork alignment were checked against the PNG composition. - The gallery supports feature comparison, enlarged viewing, arrow-key navigation, and the complete copy sequence. ## Remaining decisions 01 best preserves the current positioning. 05 explains the creation experience most directly. 03 is the strongest emotional brand alternative. Choosing between them should depend on the audience and, ideally, a store experiment rather than an invented design score. Product behavior, shipping UI, integration count, and final public-use asset approval remain production checks; see README.md. No claim of conversion uplift is made.