Design a town backward from the industrial structure it needs
A town should begin with the industrial structure it is meant to sustain, not with a fixed picture of its buildings or streets. Buildings, blocks, transport, and infrastructure are means of supporting that structure. They should be changed when demand, performance, or the industries themselves change. A rendering of a completed town can be useful, but it should not become the objective that every later decision must serve.
The starting point is the global supply chain and value chain: the demand that is expected to emerge and the activities required by national strategy. From these, a town can identify an industrial cluster and use it as an economic foundation for a human-scale, medium-sized place. This reverses the usual question. Instead of beginning with the gap between the shops and jobs that exist locally and those that are desired, the design begins with the industrial structure that the town needs to develop.
The target industrial structure is decided outside the simulator. The simulator’s role is to generate and compare many spatial arrangements that could achieve it. The result is not one master plan, but a sequence of development choices: what to build first, what to add next, and what evidence should cause the arrangement to be revised. The industrial target and the guardrails remain stable; the placement and the details of infrastructure remain adjustable.
This approach also changes the spatial objective. A town should combine open areas for exercise, walking, and nature with denser areas for commerce. Its central areas should be relatively dense, with density decreasing outward, while towns should be connected by public transport. Cars should not enter the town center. Places where people work and places where they can talk should be created within the same arrangement, rather than separated into isolated zones. The quality of an arrangement can then be judged by how well it supports the target industrial structure, how readily it can be built in stages, and how much human richness it offers through access to nature, opportunities for conversation, and limited car intrusion.
The development unit is therefore not a completed image but the placement and sequence of modules. Development proceeds by setting a target industrial structure, building initial infrastructure and modules, measuring the emergence of industries and demand, and adjusting the placement, sequence, and next module. A module may be expanded, reduced, repurposed, moved, or withdrawn. Withdrawal conditions should be defined before construction begins: demand or use may fail to reach a threshold, an occupying industry may decline, or a change in the town’s main source of earnings may require a different use. The available responses should include more than abandonment: relocation, conversion, expansion, contraction, and dismantling.
This requires modules that are easy to move, reuse, convert, or take apart. Sites should remain compatible with movement rather than being treated as permanently divided parcels. The destination for a vacated site, the equipment, and the people associated with a withdrawn activity should be considered in advance. Infrastructure should likewise be addable in small units, expandable later, and recoverable when it is not used. Each module should make it possible to measure operation and demand so that measurement can directly inform the next decision.
The practical benefit is a different risk profile for developers and investors. A small installation can produce a result quickly; if the result is poor, it can be corrected quickly. This reduces the risk of discovering only after ten or twenty years that a large, permanent development has failed. Large companies, real-estate firms, investors, and financial institutions can make decisions at each stage and see a route for withdrawal without committing the entire investment at once. The method requires corresponding changes in infrastructure development, construction modules, finance, and the treatment of movable structures under building and land-use rules.
The need for this order became clearer through a small experiment in a residential neighborhood of Zushi. Shops and places to gather had once been distributed through the neighborhood, but daily activity had become concentrated around the station. Apart from occasional school or neighborhood events, there were few places where residents could meet by chance and stop for a conversation. The experiment tested whether a nearby market could restore some of that everyday connection by using the open space that remained in the neighborhood.
Its design had three layers. Mobile vending was the point: it generated sales while building visible, face-to-face relationships. The small market was the line: it tested space, timing, movement, and the conditions under which people gathered. A public module was the surface: it would turn those relationships and observations into a physical setting whose maintenance and operating costs could circulate within the community. The underlying idea was that marketing activity could itself generate sales, while sales data could also become data about local demand.
The experiment began with temporary, low-cost arrangements. Shade, weather protection, power, signs, and a minimal boundary were added without permanent construction. The intended measures included vendor profit, reservation rates, time spent at the site, neighborhood complaints, and the time required to remove the installation. Existing vacant lots, parking areas, religious sites, and private land were considered as venues. Reservations were introduced to reduce inventory risk, and weather-dependent sales, reasons for visiting, the distribution of time spent, and removal time were recorded. Later stages would have added a roof, then power, lighting, signs, and standardized components funded partly by market revenue.
The experiment also treated participation as a gradual process. Shops would begin with reservations, so that their inventory risk was close to zero, and move gradually toward selling on the day. Several shops appearing on a fixed day could change the neighborhood’s understanding of the place into “somewhere to buy prepared food in the evening”; sporadic appearances could not create the same recognition. The platform was expected to coordinate reservations, payments, fees, announcements, venue equipment, safety and hygiene standards, relationships among shops, and negotiations with site owners.
After approximately one year, the experiment ended in September 2026. It succeeded in connecting shops and the neighborhood, and the shops continued to participate independently through their own Instagram accounts. It did not, however, become the economic core. The result led to a revision of the original bottom-up sequence: local activity alone was insufficient, and the industrial and infrastructural framework needed to come first.
That conclusion should be treated as practical evidence from one experiment, not as a general proof. The effects of the local market, its timing, and its operating structure were not separated. The experiment nevertheless provides a useful design lesson. Small local activities are valuable as tests of relationships, demand, and spatial behavior, but they should be placed on an economic and infrastructural skeleton capable of sustaining them. The town is then designed backward from the structure that must endure, while its physical form remains capable of learning, moving, and changing.