Public infrastructure is interface work at the largest scale available, and it is usually designed by default rather than on purpose.
Delhi Metro alone reportedly carries over six million passengers a day — more people than most software will ever serve. The comparison is not rhetorical: it sets the cost of an ambiguous sign.
The constraints are harder than software
A digital product can define its audience. Public space cannot. The same signage has to work for a resident in a hurry, a first-time visitor, an elderly passenger, someone with luggage, someone who cannot read the local script, and someone in distress.
The context is also loaded. People are late, carrying things, managing children, or navigating an unfamiliar city, which lowers the attention available for anything clever.
So the standard is comprehension under pressure, not elegance at rest. Every decision should be tested in the worst realistic conditions, not the best.
Design the passage, not the sign
Wayfinding fails between signs rather than at them. Map the whole passage and name the decision the traveller is making at each point.
Confirm they are in the right place before they commit.
One decision at a time, in the order it is needed.
Ticket, gate or counter: the step with the highest failure cost.
Confirmation and the route onward, including the way back.
Instrument the third. Queue length is a design output.
Four moments in a public passage: approach, orientation, transaction and exit.Placement is behavioural. One sign in the wrong position can turn an orderly flow into a bottleneck, and the correction is usually spatial rather than graphic.
Accessibility is the specification
Wheelchairs, visual and hearing impairment, cognitive difference and varying physical capability are not edge cases in public space. They are the specification, and they are also the cheapest way to improve the experience for everyone else.
Multi-script and multi-language requirements compound it. A terminal may need to work across many languages and a wide literacy range, which pushes meaning into structure, colour, iconography and sequence rather than text.
Test with the people concerned, on site. Simulation reliably underestimates how much of the environment is missing or contradictory.
What the exemplars actually did
The airports and transit systems cited as benchmarks did not win on decoration. They treated the passage as a designed sequence with a measurable outcome.
Only what is needed at that point, never the whole map.
Deliberate places to stop, which keeps flow moving elsewhere.
Waiting times treated as a metric with a target.
People positioned where the system is known to confuse.
Commercial performance followed the passenger experience, not the reverse.
Four shared characteristics of exemplary public spaces: sequenced information, relief points, measured queues and staff placement.That last point is often treated as an operations cost. It is cheaper than rebuilding a concourse and it is the fastest way to fix a known failure.
Decide what to measure
Pick one passage, count the wrong turns, time the transaction step, and ask ten people to state where they think they are going. Those three measures will tell you more than a survey.
Civic interfaces should be assessed on comprehension and completion for the least-equipped traveller, on the worst day, not on renders.
Written by the studio, from work in delivery
