Philosophy

Software should be compatible with basic human psychology. End-user tools should be minimally invasive on attention. Developer workflows should keep cognitive overhead low.

Our greatest assets are our brains. Software that spends attention for us and abuses working memory is disrespectful.

The recurring question:

When capability exists in the engine but not in the first-class UI, the product is lying about what counts.

The bug is often discoverability and product representation, not missing backend search or APIs. Power-user syntax is not a substitute for filters and settings people can find.

The same honesty rule applies to:

Consistency with existing conventions has value, but it is not absolute truth. When a convention is fundamentally backwards from human intuition and creates compounding inconvenience, it should be corrected.

The inconvenience of migration is temporary. The inconvenience of living with a backwards convention is permanent and compounding.

UX is of utmost importance in any technical system. Be kind to your future self and considerate of others' time and attention. Every developer was once a user. Pass down maintainable software that a naïve reader can still understand.

Things that behave similarly should not be presented in vastly different ways.

When a system applies rules in order, the processing-map surface should treat sequence as the top-level organization. Parallelism nests inside steps. See Processing maps and the pipeline composer essay.

Interaction-heavy prototypes should share meaning and behavior — domain IR plus a headless controller — and hand-write thin views per ecosystem. Do not chase universal UI export; compose prior art (headless machines, typed IR, swappable renderers) on purpose. See Prototype strategy and Fundamental contexts.