An idea, not a decision. Framed panes are the product: the shipped markup captured from a running serve (smoke fixture, project "Smoke Test"), edited only where the proposal changes it, at shipped CSS. The hatched block at the end is commentary. UXR and Morville are drawn as installed in the proposed catalogues; the fixture installs only Sentiment.
Five planes of user experience, from abstract to concrete — strategy, scope, structure, skeleton, surface — each dependent on the plane below it, and each with two sides: the product as functionality and the product as information.
A framework for locating where in the design stack a finding lives: the objectives and needs behind the product, the features and content it includes, how its parts are arranged, how each screen is laid out, and what the senses meet.
Is the product solving the right problem, for its users and for its makers?
Does the product include the right features and content?
How do the pieces fit together — the flows and the organisation of content?
How is each screen laid out — the interface, the navigation, the information?
What do the senses meet — colour, type, contrast, consistency?
Five planes of user experience, from abstract to concrete — strategy, scope, structure, skeleton, surface — each dependent on the plane below it, and each with two sides: the product as functionality and the product as information.
A framework for locating where in the design stack a finding lives: the objectives and needs behind the product, the features and content it includes, how its parts are arranged, how each screen is laid out, and what the senses meet.
Is the product solving the right problem, for its users and for its makers?
Does the product include the right features and content?
How do the pieces fit together — the flows and the organisation of content?
How is each screen laid out — the interface, the navigation, the information?
What do the senses meet — colour, type, contrast, consistency?
The problem. In testing, people installed from the card without knowing what they were getting, and Install starts an AutoCode run (D4: installing is applying). The card shows a title, an author and one sentence; the page shows every group and tag, the author and their sources. The decision belongs where the evidence is.
The flow. The catalogue becomes a catalogue: each card is already a button that opens its codebook (D12), so removing Install / Uninstall costs no route. Install and Uninstall already live in the page's title row; nothing moves there, it just becomes the only place. Sentiment's card already has no button today, so a button-less card is a shipped shape, not a new one.
The marker. A: a tick and "Installed" where the button was, so the eye finds it in the same place. B: a tick before the title, quieter, the word only in a tooltip (hover-only, so weaker for keyboard and screen-reader users unless it is also in the label). C: the word leads the status line, no glyph. The status phrase already differs (installed cards say "applied to…" or "none applied to…"), so the marker restates a fact the card half-carries; A makes it scannable. Not-installed cards show nothing: absence is the information.
The page line. Under the status, one sentence saying what Install does, true to the code: the first run leaves every proposal pending until it is reviewed. It is also where D4's unmet obligation could land ("the number moves to the install control"); the alternative is to put the count on the button ("Install and code 4 quotes").
CodebookV2Page.tsxis a hardcoded English literal today and would want a key at the same timeWhat would change in code. The card's button and its handlers go (
CodebookV2Browse.tsx), with the.picker-card-togglerules; one marker element and its rule arrive. The Mac Codes menu keeps its Install / Uninstall row, which acts on the focused codebook, so it is unaffected.