IELTS Academic Reading Practice Test 119

When the United States entered the Second World War, its army needed an enormous number of firing tables. These were printed booklets that told a gunner what elevation and fuse setting to use for a particular gun and shell at a particular range, so that a shell would land where intended rather than short of the target or beyond it. Each table contained hundreds of entries, and every entry required the numerical solution of a complicated equation describing the flight of a projectile through the air. The work was done by hand, by teams of mathematics graduates — many of them women, then known as 'computers' — using mechanical desk calculators. Producing a single table could keep a team busy for weeks.

The idea of calculating such tables electronically came from John Mauchly, a physicist at the University of Pennsylvania's Moore School of Electrical Engineering, who had long been interested in speeding up computation. In 1941 he met J. Presper Eckert, a young electronics engineer at the same institution, and during 1942 and 1943 the two men drafted a proposal for a machine built from vacuum tubes. The Army's Ordnance Corps agreed to fund the project in 1943, partly because the staff of its Ballistics Research Laboratory were desperate for faster calculations. The machine, officially named the Electronic Numerical Integrator and Computer, was shown to the public in February 1946.

Viewed as engineering, ENIAC was an act of nerve. A vacuum tube is a glass valve in which a small voltage on one electrode controls a much larger current between two others, and tubes fail far more often than mechanical switches. To build a fast calculating machine, ENIAC's designers used 17,468 of them, together with thousands of resistors, capacitors and relays. The completed machine weighed about 30 tons, filled a room of some 1,800 square feet and drew around 150 kilowatts of power, enough to warm the building. Its numbers were held in decimal form, in twenty accumulators each storing a ten-digit signed number, which made results easier to read than the strings of ones and zeros used by later machines.

Inside, arithmetic was performed by counting pulses, so addition was extremely fast: the machine could carry out about 5,000 additions a second, and several hundred multiplications, thousands of times faster than a person with a calculator. Because it was built from separate units wired together, it could work on several parts of a large problem at once, an early form of parallel processing.

Programming ENIAC did not resemble writing software. There was no stored list of instructions: the machine was configured by setting hundreds of switches and connecting units with cables on plugboards, a process that could take days. The people who did this work were six women — Kay McNulty, Betty Jennings, Betty Snyder, Marlyn Wescoff, Fran Bilas and Ruth Lichterman — who had been employed as human computers and had taught themselves the machine's logic from circuit diagrams, since no manual existed. Their contribution went largely unrecognised for decades: press photographs of the 1946 demonstration showed the machine, and the women were not named in the accompanying reports.

What ENIAC did first was secret. In late 1945 it was used for calculations connected with the hydrogen bomb, work that was not declassified for years. One of its best-known later tasks was in meteorology: in 1950 a group led by Jule Charney used the machine to run the first numerical weather forecasts, each of which took about twenty-four hours of computing to produce — roughly as long as the period it was predicting.

Even so, ENIAC's design was already being overtaken. Setting up a new problem took so long that it became obvious instructions should live in the same memory as data, so that a machine could be reprogrammed simply by loading a different sequence of numbers. In 1945 the mathematician John von Neumann circulated a widely read document, the First Draft of a Report on the EDVAC, which set out this stored-program principle for a successor machine being designed at the Moore School. The draft carried only von Neumann's name, a fact that angered Eckert and Mauchly and has been argued over ever since.

The first computer to run a stored program was the Manchester 'Baby', tested in June 1948; the first to be used routinely for real work was EDSAC, built at Cambridge under Maurice Wilkes and working in 1949. These machines, rather than ENIAC, established the pattern that computers still follow: a processor, a memory holding both data and instructions, and devices for input and output.

ENIAC's own end was quieter. It was formally retired in 1955, by which time transistors were beginning to replace tubes. Its legal status was settled later. Eckert and Mauchly filed a patent in 1947, which was granted in 1964, but in 1973 a court invalidated it, partly because of an earlier machine built by John Atanasoff and Clifford Berry at Iowa State College between 1939 and 1942.

Historians are therefore cautious about calling ENIAC the first computer. Britain's Colossus, built in 1943 to break codes, was electronic and earlier, but it was designed for one narrow task and its existence remained secret until the 1970s. Germany's Z3, completed in 1941, was programmable but worked with relays rather than electronic valves. What ENIAC demonstrated was not precedence but possibility: that a large electronic machine could do useful mathematics, reliably, for years.

Electricity cannot easily be stored, so for a century the grid was built on the principle that generation must match demand at every instant. Fossil and hydroelectric plants were adjusted to follow demand. Wind and solar farms cannot be adjusted in this way: their output follows the weather and the time of day. A grid-scale battery is an array of lithium-ion cells in steel containers, connected to the network through power electronics. It takes electricity off the grid when there is a surplus and returns it when there is a deficit, shifting energy from one part of the day to another.

Each cell works by moving lithium ions between two electrodes through a liquid electrolyte. When the battery discharges, ions leave the anode, usually made of graphite, and travel to the cathode, a compound of lithium and a metal oxide; electrons take the long way round through the external circuit, doing useful work as they go. Charging reverses the process. Cells based on nickel, manganese and cobalt store a great deal of energy for their weight, but they are expensive and their cobalt is mined in a few countries. Lithium iron phosphate cells, known as LFP, hold less energy per kilogram yet tolerate more charging cycles and are harder to set alight. Most large grid batteries now use LFP.

Cells are grouped into modules, modules into racks, and racks into cabinets that resemble shipping containers. A single large installation may hold well over a hundred of these containers, together with an inverter that converts the direct current of the cells into the alternating current of the grid, a transformer that raises the voltage, pumps and chillers for temperature control. All of it is coordinated by software that decides, many times a second, whether to charge or discharge and by how much.

Two figures describe a battery, and confusing them is a common error. Power, measured in megawatts, is the rate at which it can deliver energy; energy, measured in megawatt-hours, is the total amount it can store. Dividing one by the other gives the duration: a 100-megawatt battery holding 400 megawatt-hours can run at full output for four hours. Most grid batteries built so far last between one and four hours, because that is the span over which the commercial case is strongest.

Frequency regulation, keeping the alternating current at exactly fifty or sixty cycles a second, is the most demanding service such systems sell: a battery can change its output in milliseconds, far faster than a turbine can, and is paid accordingly. Batteries also replace gas-fired 'peaking' plants, which run for a few hundred hours a year at times of high demand and are expensive to operate. And they trade: charging when wholesale prices are low, often at midday when solar output is plentiful, and selling when prices rise in the evening.

The scale of the change is visible in California, where solar capacity grew rapidly in the early 2010s. Plotted on a graph, the state's net demand — the total required minus what solar panels supply — fell in the middle of the day and then climbed steeply in the evening as the sun went down. The shape resembled the profile of a duck, and the 'duck curve' became a standard term in the industry. Batteries fill the belly of the duck by charging at midday and cover part of its neck in the evening.

There is a subtler problem, too. The heaviest rotating parts of a conventional power station — the turbine and its generator — store kinetic energy, and this inertia resists sudden changes in frequency. An inverter, being electronic, has none, and a grid with a very high share of inverter-connected generation can become skittish. Engineers are therefore developing 'grid-forming' inverters, which are programmed to behave like a rotating machine, holding voltage and frequency steady and supplying the burst of current needed when a fault occurs. Such devices are now among the liveliest areas of power engineering research.

Batteries are not without costs. Their capacity fades with every cycle and with age, so a pack installed today may retain only seven or eight tenths of its original capacity after a decade. Fires are rare but severe when they happen, because a fault in one cell can heat its neighbours until the whole module burns; installation rules, spacing and suppression systems have become stricter as a result. The materials are another constraint: lithium, graphite and, in some chemistries, cobalt and nickel come from a small number of producing countries, and much of the refining is concentrated in China.

For all the excitement, lithium-ion batteries cannot solve everything. Pumped-storage hydroelectricity, in which water is pumped uphill when power is cheap and released through turbines when it is needed, still accounts for the great majority of the world's stored electricity, and the best sites have long been taken. Flow batteries, which pump liquid electrolytes from tanks through a stack of cells, separate power from energy, so a system can run for ten or twelve hours simply by building bigger tanks. They are bulky and costly but long-lived.

What has changed is cost. Lithium-ion packs became roughly ninety per cent cheaper between 2010 and 2023, and installations multiplied as a result. Batteries are now the default answer to the question of how to store electricity for a few hours, and they are likely to remain so. They are not, however, a complete answer for a grid that must run for days on end without wind or sun.

For most of the twentieth century, readers paid for their news twice. They bought the paper, and advertisers paid to reach them. When the internet arrived, publishers assumed the second stream, advertising, would simply move online, and for a while it did. They gave their journalism away in the belief that scale and traffic would bring revenue. That belief was wrong, and the reasons are now familiar: classified advertising migrated to specialist websites, display advertising was captured by search engines and social platforms that could target individuals far more precisely, and an entire generation learned to expect news for nothing.

The response has been the paywall. The New York Times, which introduced a metered system in 2011 that allowed casual readers a few free articles a month, now reports more than ten million subscriptions and earns most of its revenue from readers rather than advertisers. Other papers have followed, some with strict paywalls, others with metered ones; a few, such as the Guardian, have remained free and asked readers for voluntary contributions instead. Behind these different arrangements lies the same recognition: someone has to pay for the reporting, and the reader is the least unreliable source.

The argument in favour is straightforward. Journalism is expensive: investigations take months, foreign bureaux cost money, and courts and councils must be covered whether or not readers find them interesting. If advertising cannot pay for this work, readers must. A subscription also changes the incentives of the newsroom, because a publication paid by its audience has less reason to chase clicks with trivia and outrage, and more reason to keep the confidence of people who might cancel. On this view the paywall is not a barrier to information but the condition of producing it well.

The objection is equally serious. A paywall makes information a commodity, available in proportion to income, and the people most likely to be excluded are exactly those whose circumstances public-interest reporting exists to illuminate. Council housing, court proceedings, immigration rules and hospital failings concern everyone, but those who can afford a subscription are rarely the ones who most need to read about them. Local newspapers, which have traditionally trained reporters and held town halls to account, have been hit hardest: many have closed or retreated behind paywalls that shrink their audience to a fraction of what it was.

There is a practical obstacle as well. Households already pay for broadband, television, music and mobile telephones, and the willingness to add another monthly charge is finite. Publishers have responded by bundling: a single subscription may now include newspapers, magazines, podcasts and puzzles, and some are sold through the shops of large technology companies, which keep a share of the money. Whether such bundles deepen loyalty or merely spread thin revenue is unclear, but they are a recognition that news alone struggles to command a price.

Governments have begun to intervene. In 2021 Australia passed a law obliging the largest platforms to negotiate payments with news businesses for the links they display, and Canada followed with similar legislation in 2023; both provoked furious resistance, and in Canada the largest platform stopped carrying news links altogether. Such measures transfer money from technology companies to publishers, which helps, but they do not answer the deeper question of what journalism is for. They treat news as a product whose value can be counted in clicks rather than as a public good.

The economics of a public good are well understood. A public good is one that everyone can use without reducing what is available to others, and that is hard to withhold from those who do not pay. Lighthouses, clean air and national defence are the textbook examples. Journalism of the kind that scrutinises power has something of the same character: I benefit from living in a country where councils are reported on, whether or not I read the reports. Markets of the kind that paywalls create will always supply less of such journalism than society would choose if it could decide collectively.

None of this means that paywalls are indefensible. A reader who pays for a subscription is not withholding information from anyone; the alternative, in many cases, is not free news but no news at all, because a title that cannot pay its staff closes. The honest position is that subscriptions are a legitimate way to fund part of the press, and a poor way to fund all of it.

What would a better settlement look like? I would argue for a mixed model, deliberately built. Core reporting on the institutions that hold power — courts, councils, parliaments, hospitals — should be funded in a way that keeps it free at the point of use, whether through a public subsidy, a levy on platforms or a well-governed trust. Reviews, analysis, sport, business news and investigations, which readers value individually and will pay for, can sit behind a subscription. Philanthropy has already shown what it can do in a number of countries, though it is not a reliable substitute for a broad base of funding.

The mistake of the last twenty years was to treat the question as one of technology — print against screen, free against paid — when it was always a question of who pays, and for what. Readers will pay for journalism that is worth paying for, and some will not pay at any price. A civilised society should be unwilling to let the second group's ignorance be the price of the first group's convenience.