Shared space removes the kerb and moves the negotiation to eye contact

This is less a set of instructions about shared space streets than an argument, and it is worth saying so at the start.

Uncertainty is being used as a speed control

Shared space removes the kerb, the markings and often the signals, so no user can be sure who has priority at any given moment. The intended effect is that drivers slow down and negotiate, because the environment stops telling them they have an unambiguous right of way. This is a real mechanism and it is the same one behind traffic calming that changes geometry rather than adding instruction.

The design works by making driving slightly uncomfortable, which is why it fails wherever drivers can resolve the ambiguity in their own favour. It is worth stating plainly that the technique deliberately trades legal clarity for behavioural change, and that is a contested trade.

It depends on traffic volume more than on design quality

Below a certain flow, gaps in traffic are frequent enough that pedestrians can cross freely and drivers must yield because people are simply there. Above that flow, vehicles form a continuous stream, the informal negotiation collapses and the surface behaves like an ordinary road without crossings. The threshold is not a matter of taste, it is a function of how often a walker encounters a usable gap on the way across.

In practice, many disappointing schemes were built on streets carrying volumes that were never compatible with the principle in the first place. The honest sequence is to reduce traffic first and remove the kerb second, rather than expecting the surface treatment to do the reduction.

The kerb is a navigation aid before it is a boundary

Blind and partially sighted people use the kerb edge as a continuous tactile reference that separates the safe surface from the vehicle surface. A level surface removes that reference, and a person using a long cane can no longer tell where the footway ends without a detectable feature.

Colour contrast does not substitute for it, since the users who most need the edge are the least able to perceive a change in tone. Eye contact as a negotiation mechanism excludes anyone who cannot make it, which includes blind people, some autistic people and children of a certain age. Disability organisations in several countries have opposed level-surface schemes for exactly this reason, and that objection is about mechanism rather than preference.

Comfort and collision counts measure different things

Some shared surfaces record few injury collisions and are still avoided by older and disabled residents, who reroute or stop making the trip. A street can therefore look safe in the statistics precisely because the people most at risk have withdrawn from it. Suppressed demand of this kind is invisible in monitoring that counts users of the street rather than people who used to use it.

Measured properly, surveys of perceived safety and interviews with people who avoid a street reveal effects that collision data cannot show.

Evaluating a scheme only on recorded injuries will systematically flatter designs that discourage vulnerable users from turning up.

Hybrid designs keep a detectable edge

A common compromise keeps a continuous detectable strip, often a shallow kerb or a distinct tactile band, defining a pedestrian comfort zone free of vehicles. This preserves the navigation reference while retaining most of the informal, low-speed character the scheme was aiming for.

Across a network, bollards, planting and level changes at intervals can define the vehicle route without reinstating a conventional carriageway across the whole width. Formal crossings placed at the desire lines give anyone who wants certainty a place to get it, without forcing everybody to use them. These hybrids are less photogenic than a fully level plaza and they are what most current guidance in several countries now recommends.

What works in a dense grid may not transfer to a low-density suburb.

Judging a shared surface on site

Stand at the edge and watch vehicle speeds, because if drivers are moving at a normal urban pace the negotiation is not happening. Look for a continuous detectable feature that a cane could follow along the whole length, not just at the ends. Count how many people cross away from the marked routes, since heavy informal crossing is evidence the design is functioning as intended.

Over a decade, note who is present, and note who is not, because an absence of older people or wheelchair users on a busy street is itself a result. Ask whether the vehicles are there because they need to be, since shared space on a street full of through traffic is a decoration.

The takeaway

Take out the kerb only after taking out the traffic. In the other order it is just a wide road.

Measure the walk, not the map.