IELTS Academic Reading Practice Test 187

The vending machine is an old idea with a modern appearance. The first known description comes from Hero of Alexandria, an engineer writing in the first century AD. In a treatise on pneumatics he explained how a coin could open a valve and release a measured quantity of holy water in a temple. The coin fell onto a small pan connected to a lever; its weight tipped the lever, the valve opened and water ran out. When the coin slid off, a counterweight closed the valve. No example survives, and the device was probably never common, but the principle — payment first, delivery second — has not changed.

Commercial vending began in nineteenth-century London. In the early 1880s Percival Everitt, an inventor, patented a machine that sold postcards, and his machines were installed on railway stations and at post office premises. They were strongly built, decorated with cast iron and designed to work without an attendant. The postcard was a well-chosen product: light, uniform in size and in steady demand from travellers. Other entrepreneurs followed with machines selling stamps, matches and sweets. By the end of the century the idea had spread to Europe and the United States.

Across the Atlantic, the chewing-gum machine gave the trade a product of its own. In the late 1880s the Thomas Adams company placed small dispensers on the platforms of New York's elevated railways, where passengers waited in crowds. For a coin the machine released a single piece of gum, and the company's brightly coloured Tutti-Frutti brand became associated with the new habit. The machines were cheap to make and easy to move, and they showed that vending did not need a large cabinet or an elaborate mechanism.

The automat carried the idea further. Horn & Hardart, a Philadelphia company, opened its first automat in 1902 and brought the format to New York in 1912. The restaurant was lined with glass-fronted compartments, each holding a portion of food. A customer inserted a coin, usually a nickel, turned a handle and opened the little door to take the dish. Behind the wall, staff replenished the compartments and cooked the food, but the customer never dealt with them directly. At their height the automats served a huge volume of meals each day, and the last Horn & Hardart automat did not close until 1991.

Modern machines depend on the coin validator, a device that decides in a fraction of a second whether a disc of metal is genuine. A simple mechanical validator measures the diameter of a coin, tests its thickness in a slot and checks that its edge rolls smoothly; a magnet will also reject steel slugs. Electronic validators go further.

A coin passing through a coil changes the magnetic field around it, and the pattern of that change depends on the metal's size, shape and conductivity. The validator compares the pattern with the signatures of genuine coins held in its memory. If the coin does not match closely enough, it is returned down a chute.

Delivery is usually the work of a spiral. A motor turns a long metal coil, and the product, resting between two turns of the helix, is pushed forward until it reaches the edge of the shelf and falls into the tray below. The pitch of the coil is chosen to match the width of the product, so that a bag of crisps and a chocolate bar need different spirals. A sensor at the tray confirms that something has fallen. If nothing arrives, the machine can try again, return the money or record a fault, which is why a modern machine rarely keeps a payment without delivering anything.

Japan has the densest network of vending machines in the world. There are several million of them, roughly one for every thirty people, and in a city district a machine is rarely more than a short walk away. High land prices, low crime and a dependable electricity supply make the trade profitable, and the range of goods is unusually wide. Most machines sell drinks, hot in winter and cold in summer, but others are stocked with snacks, cigarettes or, in rural areas, fresh eggs. Some are designed to unlock automatically after a major earthquake, so that people can take free drinks and supplies.

Cashless payment has changed the economics. A machine that accepts a bank card, a transit card or a phone can report its stock and faults remotely, so staff visit only when something needs attention. It also suffers fewer coin jams and does not need to hold large sums of money in a steel box. Yet cash has not disappeared. Coins and notes still work when a network fails, and they suit people who do not have a bank account or who prefer not to leave a record of their purchases. For such customers, paying with cash protects privacy.

The vending machine is now so ordinary that it is hardly noticed. It offers a small, reliable service at any hour, without a queue and without conversation, and that anonymity is part of its appeal. The device that Hero described was a piece of temple equipment, designed to impress and to control. The modern machine is a piece of street furniture, designed to be ignored. Both, however, rest on the same bargain: the customer pays first, and the machine must then deliver.

More than a million people are thought to lose a limb each year, and many more are born without one. For most of the twentieth century a prosthetic arm or leg was a passive device: a shaped and hinged structure, often of wood or plastic, held on by straps and moved by the rest of the body. A leg of this kind can be remarkably effective, because walking is largely a matter of momentum and balance. An arm is far harder to replace, since a hand is expected to grasp, pinch, twist and release in an instant. The bionic limb is an attempt to answer that difficulty by combining a mechanical structure with electronic control. The results so far are impressive in the laboratory and uneven in daily life.

The commonest form of powered prosthesis is controlled myoelectrically. When a muscle contracts it generates a tiny electrical signal, of the order of a few thousandths of a volt, which electrodes on the surface of the skin can detect. A prosthetic hand is fitted with a socket lined with such electrodes, and a microprocessor reads the pattern of activity from the muscles that remain in the forearm. Clenching one muscle may command the hand to close and clenching another may open it. The system is intuitive enough to be learned in a few sessions, but it is crude: two electrodes give only a handful of commands, and the user must often combine them in sequence, for example to rotate the wrist before gripping.

Newer systems extract far more information from the same signals. Rather than assigning one muscle to one movement, pattern-recognition software compares the whole electrical picture with a library gathered during training, and infers which of a dozen or more movements the user intends. Machine learning makes this practical, since the mapping from signals to intentions differs from person to person and shifts as muscles tire or the socket is worn differently. In laboratory tests, trained users can control individual fingers of a powered hand. In everyday life the accuracy is lower, because sweat, changes in posture and the pressure of the socket all distort the signal.

A more radical approach goes inside the body. In targeted muscle reinnervation, developed by the surgeon and engineer Todd Kuiken in Chicago in the early 2000s, nerves that once controlled the amputated limb are surgically rerouted to spare muscles in the chest or upper arm. When the patient thinks about moving a missing hand, these muscles contract, and their signals are far stronger and more distinct than anything that can be recorded from the surface of a residual limb. The procedure can also restore a degree of feeling: as the skin over the rerouted nerves is touched, patients report sensations that appear to come from the missing hand, including the location and quality of the touch.

Restoring sensation is now a field in its own right. Without it, a user must watch the prosthesis to know how hard it is gripping, and a cup may be crushed or may slip. Researchers are testing electrodes implanted around or inside peripheral nerves, which can both record motor commands and deliver small currents that the brain interprets as touch. In trials, volunteers have reported feeling the shape and stiffness of objects, and have gripped them more gently as a result. The challenge is durability: nerves tolerate implanted electrodes less well than the brain does, and the number of distinct sensations that can be produced reliably is still small.

Even a perfect control system is limited by how the limb attaches to the body. A socket must grip the residual limb tightly enough not to twist, yet loosely enough not to damage the skin; in hot weather it becomes uncomfortable, and repeated pressure produces sores. One alternative is osseointegration, in which a titanium implant is fixed into the bone of the residual limb and emerges through the skin, so that the prosthesis can be attached directly to it. The method, which grew out of research in Sweden by Per-Ingvar Brånemark, gives a more secure and comfortable fit, but the point where the implant leaves the body is a permanent route for infection.

Power remains a stubborn constraint. Motors, gearboxes and batteries add mass exactly where mass is least welcome, and an arm that is tiring to wear will be left in a cupboard. Batteries must be recharged daily and are vulnerable to water. Cost is an even greater barrier: a multi-articulating hand such as the i-LIMB, launched in 2007, cost tens of thousands of pounds, and sophisticated arms developed with military funding, including one approved by the American regulator in 2014, are available to very few of those who need them. Most amputees in the world receive no powered prosthesis at all.

The most ambitious research bypasses the muscles altogether. Electrodes implanted in the brain can record the activity of groups of neurons that encode movement, and software can translate that activity into commands for a robotic arm. In a widely reported experiment published in 2006, a man paralysed by a spinal injury used such a system to move a computer cursor and to open and close a prosthetic hand. Brain implants of this kind are still experimental, and they raise difficult questions about long-term reliability and consent. For the foreseeable future, most bionic limbs will continue to listen to muscles rather than to thoughts.

Few industries have been rearranged as completely, and as quickly, as recorded music. In 1999 the business was at its height, selling compact discs at prices that made it one of the most profitable corners of the entertainment world. Fifteen years later its global revenue had fallen by roughly 40 per cent. The recovery that followed, driven almost entirely by subscription streaming, has been celebrated by investors and, more cautiously, by musicians. Yet the argument about streaming is often conducted as though the only question were whether it has rescued the industry. The more interesting question is whom it pays, and how.

The collapse had little to do with streaming, which barely existed when it began. It was caused by the arrival of cheap broadband and file-sharing software, which made copying free and, for a generation of listeners, normal. The best-known sharing service of the era was forced to close after legal action in 2001, but the habit survived. Compact discs had also encouraged a business model based on selling albums at a high price to people who wanted one or two songs. Once a single track could be downloaded in a minute, that model had no future. The industry responded slowly, and partly by suing its own customers.

Streaming offered a way out. The Swedish service Spotify, launched in 2008 and available in the United States from 2011, did not so much sell music as license access to it, paying rights holders a share of subscription and advertising income. Listeners responded enthusiastically: tens of millions of recordings became available at the touch of a screen, and the habit of paying for music, abandoned by many during the file-sharing years, returned in a new form. Labels, for their part, were initially reluctant, having seen their business undermined by unlicensed copying, and they licensed their catalogues only gradually. By 2022 global revenue had recovered to roughly its 1999 level.

There are genuine gains that the old business never offered. Back catalogues have been revived, and performers who made their recordings decades ago now earn from listeners in countries where their records were never stocked. An artist working far from the traditional centres of the industry can be heard anywhere. Playlists, too, have replaced radio as the way many people encounter new music. For listeners, the range of music available is without historical precedent, and the cost of exploring it is close to nothing. It is worth saying plainly that this is an improvement, not a loss.

The criticism concerns distribution. In the standard model, all the money a service collects from subscriptions is pooled and divided according to each track's share of the total number of plays. This is known as pro-rata payment. Its effect is that a recording played once by a devoted admirer, who chose it deliberately, earns exactly as much as one played in the background of a shop. Money follows volume rather than attention, which rewards music that is long, repetitive or placed on prominent playlists, and gives almost nothing to an artist whose audience is small but committed.

A second complaint is about concentration. The three major record labels account for roughly two-thirds of global recorded music revenue, and the largest share of streaming income flows to a small number of very popular recordings. Meanwhile the removal of the old bottlenecks has produced an extraordinary flood: around a hundred thousand new recordings are uploaded every day, and more than a hundred million now sit in the catalogues. Distribution has never been easier, but being heard has never been harder.

One proposed remedy is to change the accounting. Under user-centric payment, the subscription paid by each listener would be divided only among the artists that listener actually played. A dedicated fan of an obscure band would then support that band directly, instead of sending most of the money to whoever topped the charts. Several smaller services have experimented with the idea, and one large platform has announced that it will move in that direction. The arithmetic is not complicated, and modern billing systems make it feasible. Its supporters argue that it aligns payment with choice and reduces the incentive to inflate play counts artificially.

Scepticism is warranted. User-centric accounting does nothing for the many listeners whose tastes are mainstream, and it is administratively awkward, since it requires a separate calculation for every subscriber. More fundamentally, the split between artists and the companies that own their recordings is set by contracts, not by the payment model. A parliamentary committee in the United Kingdom concluded in 2021 that performers received pitiful returns from streaming, and the remedy it proposed was not a different formula but a fairer allocation of power between creators and intermediaries.

None of this amounts to an argument for turning back. The alternative to streaming is not the compact disc; it is the file-sharing era, when musicians were paid nothing at all and listeners had no legal route to the music they loved. Streaming has restored revenue, widened access and given old recordings a second life. What it has not done is distribute its gains fairly. That is a problem of regulation, contracts and accounting, and it will not be solved by nostalgia. It requires greater transparency about how money is divided, and a fairer balance of power between those who make recordings and those who own them.