After an interruption, operators rebuild their picture of the plant by returning to one stable overview
Aliases: overview display as an orientation anchor · industrial human factors
What it is
An overview display can act as an orientation anchor: after an interruption, a handover, or a deep diagnostic excursion, an operator returns to one stable, global representation to reconstruct system structure, principal states, and unresolved anomalies. Its value is not in showing more information — it is in giving the recovery process a fixed starting point, so that wherever the interruption happened, this one screen is always where reorientation begins.
Why it happens
Recovering situation awareness means rebuilding three things at once: where am I, what is the process doing now, and what happened while I was away. A layout that stays spatially stable lets the operator reuse existing spatial memory instead of searching from scratch; summary states point attention to what actually needs it; an event trace fills in changes that happened but were never directly observed. Miss any one of these and recovery falls back on slower, item-by-item checking.
The decisive factor is not whether an overview exists, but whether its elements stay put. If the overview is laid out by fixed topology — component positions track the actual process or plant geography, so the same device is always in the same place before and after the interruption — the operator's spatial memory is still valid on return, and reorientation is fast. If instead the overview is a dynamically re-sorted list, say active alarms re-ranked by priority or recency so row positions keep changing, the "where was it last" memory an operator relies on is invalidated outright, and the anchor function disappears even though the display is still technically an overview — because the anchor depends on position staying fixed, not on the information being complete. This is the condition that decides whether a given overview can actually orient someone or not.
Studying it
Interrupt simulated work and have participants resume from varied entry points, measuring time to the next correct action, the number of missed changes, and the review path taken. Varying layout stability and the presence of an event history as separate factors distinguishes true orientation recovery from mere navigation speed — if two conditions show similar resumption time but very different counts of missed changes, the difference is happening at the perceptual-reconstruction step, not at operator proficiency.
This line of work follows the same logic as resumption-lag measurement in the interruption-resumption literature: the time between the end of an interruption and the first correct action is a more direct index of recovery quality than asking participants afterward how much they remember.
Where it stops holding
An overview's aggregation rule can produce an all-green illusion: a severe local anomaly averaged into a summary state can be invisible on the overview, so reading "normal" off the overview can itself be the wrong call. No single screen can represent every dependency in a large system either — the overview's job is orientation and traceability, not replacing subsystem displays for fine-grained judgment.
The anchor effect also depends on the operator already holding a stable mental model of that topology. For someone new to the role who has not yet built the mapping between the display and the real plant, the overview is just a set of icons in positions with no prior meaning — there is no spatial memory to reactivate, so a fixed layout offers little advantage over a dynamic list until that mapping has been learned through other means.
Applying it
Fix the position of the core topology and key indicators on the overview; do not let alarm-priority changes reshuffle where things sit. Mark new events since the interruption, the current selection, and the data's timestamp directly on the display, so the operator can tell at a glance what was already there before they left and what just happened.
Make both shift handover and interruption recovery start from this one screen as a fixed procedural entry point, rather than letting an operator pick up work directly from an arbitrary detail page.
How to check: run an interruption-resumption test in which an operator steps away from the console and is then asked, within a limited time, to state what changed, the current risk, and the next action, using only the overview. Compare completeness and time against both a stable version with event history and a dynamically re-sorted version with no history — only that contrast confirms whether the anchor design is doing anything.