Change Details

Evaluate where the carousel code should live and how data flows between components. Produce a recommendation that the team can act on for the carousel port task. This task should be completed before other carousel subtasks are picked up. === Questions to answer **Where should the code live?** - MinervaNeue is the leading candidate (skin-level presentation, mobile-only). Could it alternatively live as its own module in MMV? What are the tradeoffs? - What code from the prototype needs to be ported? (`thumbExtractor.js`, `excludedImageSelectors.js`, carousel UI component, etc.) - We don't want this in MobileFrontend **How do thumbnails get from MMV to the carousel?** - Should MMV's bootstrap fire a client-side hook with discovered, non-excluded thumbnails that Minerva consumes? - What should the shape of that hook be? (image URLs, file titles, alt text, dimensions?) - Can we guarantee MMV's discovery runs before Minerva needs the data? **Server-side container injection** - What server-side check is sufficient for Minerva to confidently inject a placeholder? (e.g. raw image count in page HTML, basic `noviewer` filtering, or something simpler?) - How do we handle the edge case where the server reserves space but client-side filtering discovers fewer than 3 qualifying images? - Is there a risk of the empty placeholder being visible for too long on slow connections? If so, how do we mitigate (skeleton UI, animation, graceful collapse)? **Feature gating** - What feature flag mechanism should we use for incremental rollout? URL parameter, MediaWiki config variable, or both? - Should this be the same flag as the beta viewer (T417675), or independent? **Display rules** - Only articles with 3+ qualifying images. Does the server-side check or the client-side logic enforce this threshold? - How should the carousel connect to the beta viewer? Confirm that navigating to `#/media/File:...` on tap is the right approach. === Deliverable A brief written recommendation (can be a comment on this task or a short document) covering the above questions, with enough detail for the port task to proceed.

Evaluate where the carousel code should live and how data flows between components. Produce a recommendation that the team can act on for the carousel port task. This task should be completed before other carousel subtasks are picked up. === Questions to answer **Where should the code live?** - MinervaNeue is the leading candidate (skin-level presentation, mobile-only). Could it alternatively live as its own module in MMV? What are the tradeoffs? - What code from the prototype needs to be ported? (`thumbExtractor.js`, `excludedImageSelectors.js`, carousel UI component, etc.) - We don't want this in MobileFrontend **How do thumbnails get from MMV to the carousel?** - Should MMV's bootstrap fire a client-side hook with discovered, non-excluded thumbnails that Minerva consumes? - What should the shape of that hook be? (image URLs, file titles, alt text, dimensions?) - Can we guarantee MMV's discovery runs before Minerva needs the data? **Server-side container injection** - What server-side check is sufficient for Minerva to confidently inject a placeholder? (e.g. raw image count in page HTML, basic `noviewer` filtering, or something simpler?) - How do we handle the edge case where the server reserves space but client-side filtering discovers fewer than 3 qualifying images? - Is there a risk of the empty placeholder being visible for too long on slow connections? If so, how do we mitigate (skeleton UI, animation, graceful collapse)? **Feature gating** - What feature flag mechanism should we use for incremental rollout? URL parameter, MediaWiki config variable, or both? - Should this be the same flag as the beta viewer (T417675), or independent? **Display rules** - Only articles with 3+ qualifying images. Does the server-side check or the client-side logic enforce this threshold? - How should the carousel connect to the beta viewer? Confirm that navigating to `#/media/File:...` on tap is the right approach. === Deliverable A brief written recommendation (can be a comment on this task or a short document) covering the above questions, with enough detail for the port task to proceed. --- Spike document: https://docs.google.com/document/d/1z5-BXNUXZu3AZ1sNErziFvefJPSjrXfe40kxYpxxTfA/edit?tab=t.0#heading=h.glznqc2yzi79

Evaluate where the carousel code should live and how data flows between components. Produce a recommendation that the team can act on for the carousel port task. This task should be completed before other carousel subtasks are picked up. === Questions to answer **Where should the code live?** - MinervaNeue is the leading candidate (skin-level presentation, mobile-only). Could it alternatively live as its own module in MMV? What are the tradeoffs? - What code from the prototype needs to be ported? (`thumbExtractor.js`, `excludedImageSelectors.js`, carousel UI component, etc.) - We don't want this in MobileFrontend **How do thumbnails get from MMV to the carousel?** - Should MMV's bootstrap fire a client-side hook with discovered, non-excluded thumbnails that Minerva consumes? - What should the shape of that hook be? (image URLs, file titles, alt text, dimensions?) - Can we guarantee MMV's discovery runs before Minerva needs the data? **Server-side container injection** - What server-side check is sufficient for Minerva to confidently inject a placeholder? (e.g. raw image count in page HTML, basic `noviewer` filtering, or something simpler?) - How do we handle the edge case where the server reserves space but client-side filtering discovers fewer than 3 qualifying images? - Is there a risk of the empty placeholder being visible for too long on slow connections? If so, how do we mitigate (skeleton UI, animation, graceful collapse)? **Feature gating** - What feature flag mechanism should we use for incremental rollout? URL parameter, MediaWiki config variable, or both? - Should this be the same flag as the beta viewer (T417675), or independent? **Display rules** - Only articles with 3+ qualifying images. Does the server-side check or the client-side logic enforce this threshold? - How should the carousel connect to the beta viewer? Confirm that navigating to `#/media/File:...` on tap is the right approach. === Deliverable A brief written recommendation (can be a comment on this task or a short document) covering the above questions, with enough detail for the port task to proceed. --- Spike document: https://docs.google.com/document/d/1z5-BXNUXZu3AZ1sNErziFvefJPSjrXfe40kxYpxxTfA/edit?tab=t.0#heading=h.glznqc2yzi79