Intuitive UX - locality, controls, squint test

Digested from Erik D. Kennedy, 4 Rules for Intuitive UX and The 3 Laws of Locality. Paraphrased in our vocabulary.

1. The three laws of locality (where controls live)

A control sits where it effects change. The user's eye goes to the place that will change; the control must already be there. "Add" lives at the end of the list, folder actions on the folder, sign-out next to the account avatar. Applies to conceptual groupings too, not just data rows.
A control that acts on a whole area sits above that area. Refresh-inbox above the threads; app-wide search above page-level search. Recursive: areas nest, and the hierarchy of position mirrors the hierarchy of scope.
The farther a control is from its effect, the more it must pop. If you must displace a control (thumb reach, corner actions), pay the tax in visual weight - floating compose button, strong color. The best fix remains putting it back where it belongs.

2. ABD - anything but dropdowns

Dropdowns hide options, cost extra taps, and collapse under long lists. Replace by option count:

2 options → segmented buttons (equal use), checkbox (off-by-default, applies on save), switch (immediate effect).
2–5 options → vertical segmented buttons (long labels), radios (options carry sub-content), cards, or literal visual choices (show the thing).
Many options → typeahead search.
Dates → near-future clustered dates get a calendar; wide-spread dates get a text field; never a dropdown trio.
Counts → stepper (small quantities dominate).
Dropdown tolerated only when: default rarely changed, ≤3 options, desktop-only.

3. The squint test

Squint: the most important thing must catch the eye first, the least important last. Identify the most important thing by frequency of use (per 100 visits, what happens ~95 times?), emphasize it, and demote or delete the rest. His case study cut a form's visible elements from 10 to 5.

4. Teach by example

Concrete examples beat descriptions: real use cases on the landing, pre-filled sample data on first load, inline help at the point of doubt, galleries of what others made. A sentence of positioning only lands after the examples.

Delta vs our system

CONFIRMS - Squint test = our 5-second pre-verbal test. Same rule, independently formulated: the page must communicate before it is read. Kennedy gives it an operational trigger (frequency count → most important thing) that sharpens ours. - "Emphasize one, demote the rest" = scarce emphasis. - "The page's heart owns ≥50% of the first screen" is our stronger, quantified version of his squint priority.

ADDS - The three laws of locality - we have no control-placement rule at all. Our system governs reading (type, space, color); it says nothing about where a button, filter, or action goes on a board or form. Adopt verbatim as three checks. - ABD control-selection table. We have no rule for choosing input controls. The by-option-count table is directly usable in eR form reviews and our dashboards. - Teach by example / sample data on first load. Nothing in our system covers empty states and first-run learning; our Jamaica "∅ empty registry" gesture is a seed, this generalizes it.

TENSION - None substantive. This is the complementary half our system lacks: ours governs the reading surface, Kennedy's governs the interaction surface.