A. For almost a century, the elevated reservoir above the town of Marston supplied drinking water to its residents. Built in 1912, it consisted of a broad, covered tank surrounded by a narrow maintenance path. Water arrived from a treatment works in the hills and travelled downhill through the town's pipes without additional pumping. By 2008, however, changes in water-quality requirements had made its inspection chambers difficult to maintain. A modern tank closer to the treatment works took over its function. The old reservoir was drained, its gates were locked, and its future became a subject of unusually persistent local debate. Its useful working life had ended, but the structure itself remained sound.

B. The first proposal was straightforward: demolish the tank and sell the land for housing. Supporters pointed to the cost of maintaining an empty structure and the need for more homes near the centre. Yet the council's preliminary investigation revealed an unexpected complication. The tank's walls helped stabilise the slope, and removing them would require expensive new retaining works. A second proposal, to turn the site into a museum, was also examined. Interviews suggested that residents liked the idea in principle but would rarely make a special journey to visit it. The council therefore invited suggestions that would give the place a regular role in everyday life rather than relying on occasional cultural tourism.

C. A landscape architect, Elena Voss, proposed a sheltered public garden above the tank and a rainwater store inside it. Crucially, she did not suggest returning untreated water to the drinking-water network. Rain collected from the garden paths would instead irrigate planted areas during dry weather. This separation reduced the complexity of the design. The garden's entrances would connect two streets previously divided by a steep detour, making the site useful even to people with little interest in gardening. Voss argued that a route people already needed was more likely to keep the space active than a collection of attractive plants alone. Her proposal won support after residents tested the route during a weekend opening.

D. Construction began with a year of observation rather than immediate planting. Engineers installed sensors to measure small movements in the roof, while volunteers recorded where water gathered after storms. The records showed that the western corner dried more slowly than the rest of the site. As a result, the planting plan was changed: moisture-loving species were placed there, and drought-tolerant plants occupied the sunnier eastern edge. The original drawings had treated both areas alike. This modest alteration later became a favourite example in the project's public talks. It demonstrated that apparently uniform pieces of infrastructure can contain important local differences, and that collecting data before making permanent decisions can save later expense.

E. The garden opened in 2016. Its managers initially expected retired residents to provide most of the volunteer labour, because weekday mornings seemed the most convenient time for maintenance. In practice, the largest volunteer group consisted of employees from nearby shops who came after closing time. Maintenance sessions were moved to early evening, and the council provided a secure cupboard so tools did not have to be carried up the hill. Visitors were free to sit, walk or eat lunch without joining any activity. Entry was free, and individual garden plots were not rented out. The arrangement deliberately avoided making membership a condition of enjoying a space that had previously belonged to the town's water service.

F. An independent review compared the garden with two conventional parks of similar size. It counted visits at fixed times during spring and autumn and interviewed people on the connecting streets. Visits to the reservoir site were concentrated around commuting hours, whereas the other parks were busiest on weekend afternoons. The reviewers cautioned against treating total visitor numbers as a complete measure of success: a five-minute crossing could be valuable even if it did not resemble a leisure visit. The study did not measure changes in surrounding house prices. It did record complaints about evening noise, which led to the relocation of a seating area away from neighbouring bedrooms. The managers published both the favourable findings and the complaints.

G. The project did not provide a universal recipe for converting obsolete reservoirs. Its financial case depended partly on avoiding the slope-repair costs associated with demolition, a saving that another site might not offer. Nor could every tank support the weight of a public garden. Nevertheless, the Marston experience changed the questions local officials asked about redundant infrastructure. They began by examining what a structure already did for its surroundings, including functions that were not part of its original purpose. The reservoir had stored drinking water, supported a hillside and separated two streets. Recognising all three roles made it possible to imagine a useful successor without pretending that the old engineering problem and the new public-space problem were identical.

A. At the Lanton Institute of Design, first-year students once began their training by learning the names of dozens of visual principles. They could define contrast, rhythm and balance in an examination, yet many overlooked these qualities when looking at an unfamiliar street or room. Lecturer Farah Nadir suspected that the problem lay not in the amount of information students received but in the order in which they received it. Definitions supplied too early could become labels pasted onto whatever happened to be most obvious. She replaced one introductory lecture with an apparently simple assignment: visit the same public square on three different days and record ten observations on each visit. No specialist terminology was required.

B. The first notebooks were full of broad judgements. Students called the square welcoming, untidy or busy, but seldom explained what had produced these impressions. Nadir then asked them to distinguish description from interpretation. 'A delivery van stopped beside the eastern gate at nine' was a description. 'The gate is badly designed' was an interpretation that needed additional evidence. This distinction was not intended to exclude opinions. Rather, it made opinions easier to discuss because their foundations could be examined. On later visits, students began noting the width of routes, the direction of shadows and the pauses people made before choosing a seat. Their language became more specific even before their drawings became more accomplished.

C. To investigate the effect systematically, Nadir worked with educational researcher Tomas Bell. One group received a detailed checklist before visiting the square; another received only a blank notebook and the instruction to observe. Both groups spent equal amounts of time on site and later attended the same seminar. The checklist group recorded more of the features named on its sheet, including signs and lighting. The blank-notebook group recorded a wider range of unlisted events, such as children adapting a low wall into a game. Neither group was simply better. A checklist improved consistency when observers needed comparable information, while the blank page made unexpected features more likely to enter the record. The researchers regarded the tools as serving different purposes.

D. A second experiment concerned memory. Students watched a short film of people entering a building and then wrote accounts of what they had seen. Some took notes while watching, while others wrote only after the film ended. Immediate note-takers preserved more details about the order of events, but their accounts were not automatically more accurate in every respect. Several confidently recorded an umbrella that was actually a rolled poster. When all students watched the film again, those who had left space beside their first notes were more likely to add corrections instead of silently replacing the original account. Bell suggested that a visible record of revision could make uncertainty easier to manage, since changing one's mind no longer meant concealing an earlier mistake.

E. These findings led to a notebook format with three columns: observation, possible explanation and further question. Students were encouraged to leave the second column empty when they lacked evidence. A note that sunlight reached a bench at noon did not have to become an immediate claim about why people chose it. The third column could instead propose a later visit at a different time. The format slowed down some early assignments, which initially worried the teaching team. By the end of term, however, students usually needed fewer revisions to their design proposals because they could trace each recommendation back to observations. The team did not claim that three columns were uniquely effective; they valued the separation of roles that the columns made visible.

F. Digital tools introduced both opportunities and distractions. Photographs preserved details that could be checked later, and time stamps made sequences easier to reconstruct. Yet a large image collection sometimes encouraged students to postpone looking carefully until they returned indoors. Nadir asked students to select only two photographs from each visit and explain what each contributed to the written account. This limit was a teaching device, not an argument that photographs were inherently inferior to sketches. The institute also lent cameras to students who did not own suitable phones. Differences in equipment were therefore less likely to be confused with differences in observational ability, although the researchers acknowledged that students arrived with unequal experience of visual work.

G. The notebook method later spread to courses in engineering and social research. Its appeal lay partly in its modesty: it did not promise that beginners would see everything, or that observation could ever be entirely free of assumptions. It offered a way to make the movement from noticing to explaining more deliberate. Bell warned against grading notebooks solely by the number of entries, since students could satisfy such a rule with repetitive or trivial statements. A useful notebook sometimes contained a single well-developed question. For Nadir, the strongest evidence of progress was a student returning to a familiar place and discovering something that an earlier confident judgement had prevented them from seeing. Expertise, in this account, included learning when to look again.

A. When public services adopt digital booking systems, their success is often described in terms of reduced waiting time. An appointment replaces a queue, and a confirmation message replaces an uncertain journey. These are real gains. Yet the language of convenience can conceal a less visible question: whose time has become easier to organise, and whose has become harder? A system may reduce the average wait while making a small number of people's schedules much less predictable. My concern is not that appointments are undesirable. It is that we too readily treat a single measure of efficiency as a complete account of what a service provides. Time is not merely a quantity that can be saved; it is also something people must coordinate with others.

B. Consider the fictional borough of East Mere, which changed its housing advice service from a morning walk-in queue to online appointments. Administrative records showed a sharp reduction in time spent in the waiting room. Staff could prepare relevant documents before each meeting, and many residents welcomed the change. A follow-up survey, however, revealed that some shift workers could not know their working hours far enough ahead to select a reliable slot. Others depended on relatives for transport. The old queue had been inconvenient, but it allowed these residents to choose a day at short notice. Their difficulties did not appear in the waiting-room statistics because many never reached the waiting room at all.

C. In response, the borough introduced a small number of same-day appointments that could be requested by telephone as well as online. This did not restore the old queue or solve every problem. It did make the service more usable for people with uncertain schedules. Crucially, managers assessed the change by examining completed consultations and abandoned attempts to obtain help, rather than simply counting booked slots. The distinction matters. A full calendar can be evidence of demand, but it can also conceal repeated failed attempts by the same people. An empty slot may represent wasted capacity, or it may provide necessary room for an urgent case that cannot be planned days in advance.

D. Some commentators suggest that better prediction will eventually eliminate such tensions. If software can forecast demand accurately enough, they argue, services can offer exactly the right number of appointments at exactly the right times. Prediction can certainly help with recurring patterns, such as seasonal applications or school-holiday travel. But forecasting total demand is different from knowing the circumstances of each person who needs help. A model can correctly predict that one hundred residents will seek advice next Tuesday without identifying which of them will lose access to transport on Monday evening. More data may improve a forecast while leaving this difference intact. Uncertainty is not always a temporary defect that a larger database will remove.

E. Another proposal is to let users exchange appointment slots through an online marketplace. In services where everyone has similar information and freedom to rearrange commitments, this could be useful. In many public services, those conditions do not hold. A person who checks a phone throughout the day may obtain a convenient slot more easily than a person whose job prevents frequent access. The apparent freedom to trade can therefore reward existing flexibility. This is not an argument against all exchanges; a simple cancellation list may benefit many users. It is an argument for examining who can participate before declaring a mechanism fair merely because its rules are the same for everyone.

F. Service designer Miriam Holt proposes what she calls a flexibility budget. Managers reserve part of their capacity for cases that cannot fit a standard timetable and make the purpose of that reserve explicit. The budget has a cost: on a quiet day, some reserved capacity may go unused. Holt argues that this should be evaluated in the same way as other forms of resilience. A spare pump is not necessarily a mistake because the main pump worked today. Equally, reserve appointments should not become an excuse to ignore persistent inefficiency. Managers need evidence about how often the reserve is used, whom it helps and whether its size remains appropriate. Flexibility becomes a managed commitment rather than an invisible favour granted by an individual employee.

G. The broader issue is how institutions describe improvement. If every reform is evaluated only by average waiting time, organisations have an incentive to make difficult cases disappear from the measurement. A more informative account would report the distribution of delays, the number of people who abandon an attempt to obtain help and the reliability of the promises made to users. No single arrangement will suit every service: an emergency clinic and a museum tour face different demands. But both can ask whose commitments their timetable assumes to be adjustable. Digital booking is valuable when it expands the ways people can reach a service. It deserves closer scrutiny when it merely transfers the burden of uncertainty from the organisation to people least able to absorb it.