Home > FAQs > The 3-axis framework

What is the 3-axis framework for evaluating interfaces?

The 3-axis framework evaluates a HCI (Human-to-Computer Interface) on three costs borrowed from fighting-game design: start-up frames, active window duration, and end-lag. Together they measure how much of the user's time and attention the interface lays claim to, before, during, and after use. It comes from the HCI essay on this site.

The three axes

Start-up frames: how long the user waits, or how many prerequisite actions they take, before they can engage the interface at all. The essay's key correction: start-up frames count from where the user already is, not from where the interface sits. A button on another screen has high start-up frames even if its click latency is zero; a keyboard shortcut has near-zero start-up frames if the hands are already typing.
Active window duration: how long the process takes from initiation to result. A blocking modal is the canonical bad case. Background processing softens the perceived cost but doesn't shorten the real one: if the user needs the result to proceed, they're still waiting.
End-lag: the effort required to return to the pre-interface state. Modals that demand dismissal have it; tools that pop the user out of focus have it. The ideal is zero: the user finishes the action and is already where they want to be next.

Why these three

The essay's test: a HCI that scores low on all three feels fluent; scoring high on any single axis breaks flow. Its summary line: a great HCI takes nothing from the user that the work itself didn't already demand. The heuristic checklists design schools teach measure whether an interface is usable; these axes measure whether using it preserves fluency.

The fighting-game lineage

The vocabulary comes from a domain where players spent thirty-plus years refining language for exactly this property: how committal is each move, and what does it cost the player who throws it? The essay offers the borrowing itself as a design lesson in cross-domain pollination, and OpenCues' design (typeable triggers, no popups, in-place results) is the framework applied.