Fundamental contexts for portable UI
Status: discussion stub — the long-form brainstorm lives on the blog.
Thesis
Portability may not require one mega-spec. Interactive systems share a small set of contexts — structure/semantics, placement, events, 2D (and sometimes 3D) surfaces, zoom/navigation — that can be specified sharply while host chrome stays native.
Sharable vs native
| Context | Can specify | Stays native |
|---|---|---|
Structure / semantics |
Domain IR, validation, evaluation |
Copy, empty states, delight |
Events / interactivity |
Controller machines, roles, shortcuts |
Platform a11y idioms beyond baselines |
Layout / placement |
Renderer capability profiles |
Theme tokens, density, design-system chrome |
2D / 3D presentation |
Renderer choice (DOM list vs canvas) |
Asset and animation polish |
Zoom / navigation |
Controller navigation modes per zoom level |
Gestures tied to host OS |
Zoom is not a skin — it changes the interaction paradigm (focus context, level-of-detail, spatial memory).
Stack (intentional composition)
Prior art tackled parts (Zag, Mitosis, Web Components, XState, node editors, QML, Power Automate). HCI Nerdz prototypes stack them:
-
Same IR for DOM and canvas — see pipeline-composer
Read next
-
Blog (exploratory): Fundamental contexts for portable UI
-
Strategy (pattern): Share the IR, not the pixels