IELTS Academic Reading Practice Test 118

Before the nineteenth century, weaving a patterned fabric was a slow craft. A plain cloth requires only that the weft, the thread carried by the shuttle, pass alternately over and under the warp threads, which run lengthways. A figured silk is different: selected warp threads must be raised and the rest left down at every pass of the shuttle, so that the coloured weft appears only where the design needs it. In Lyon, the centre of the French silk trade, the lifting of the right cords fell to a drawboy, often a child, who sat above the loom and worked on the weaver's command. A complex design might require thousands of such decisions, and one error spoiled the cloth.

Inventors had been trying to mechanise the work for most of the eighteenth century. In 1725 Basile Bouchon, a French silk weaver, replaced the drawboy with a roll of perforated paper through which needles could pass. Three years later Jean-Baptiste Falcon joined stiff cards into an endless chain, so that the order of operations could be altered more easily. In the 1740s Jacques de Vaucanson, better known for his mechanical automata, built a loom in which a perforated metal cylinder controlled the hooks. None of these machines was reliable enough for daily use, but together they established a principle: that a pattern could be stored in a physical medium and read by the machine itself.

Joseph Marie Jacquard was born in Lyon in 1752, the son of a master weaver. He tried bookbinding, wire-drawing and several other trades before returning to the loom, and he knew Vaucanson's machine, which he had examined in a museum. In 1801 he showed a mechanical loom at the Paris industrial exhibition, and in 1804 he patented the attachment that carries his name. Napoleon, eager to rebuild French industry after the Revolution, ordered it into production. In 1806 the government bought the rights and declared the invention public property, granting Jacquard a pension and a small sum for every loom that used it.

The attachment sits at the top of an ordinary hand loom. Its central component is a square prism, which turns to present a new face at every pick, the term for one pass of the shuttle. A chain of cards is laced into an endless loop and drawn over the prism, one card for each pick.

Behind it a row of horizontal needles presses forward against the card. Where the card is solid a needle is stopped; where there is a hole the needle passes through, and its far end releases a hook, which lifts the warp thread attached to it. The holes therefore decide which threads rise, and the woven pattern is the sum of those decisions, card after card.

The effect on production was dramatic. One weaver, assisted by the machine, could produce figured silks that had previously needed two, and far more elaborate designs became practicable. A celebrated demonstration piece, a woven silk portrait of Jacquard, was made in 1839 and used around 24,000 cards. Not everyone welcomed the change: silk workers in Lyon, fearing for their livelihoods, attacked and smashed machines, and troops were sent to the city. In the longer run, however, work in the industry increased. By 1812 there were some 11,000 Jacquard looms in France, and by mid-century the attachment was in use in weaving centres across Europe and North America.

Among those impressed was Charles Babbage, the English mathematician and inventor. Babbage owned one of the woven portraits and saw at once what the cards implied: if a machine could follow a sequence of instructions recorded in holes, it could in principle carry out any calculation that could be broken into steps. He adopted the same principle for the Analytical Engine, the general-purpose computer he began designing in the 1830s, and Ada Lovelace devoted part of her celebrated commentary of 1843 to explaining how the cards would direct its operations.

The next large-scale use of cards arose from a different problem. The United States census of 1880 took years to process by hand, and the constitution requires a count every ten years. Herman Hollerith, an engineer who had worked on the 1880 count, devised a system in which each person's details were recorded as holes in a card and read by electrical contacts that closed when a wire passed through a hole. The 1890 census was processed in about a year. The company Hollerith founded in 1896 later formed part of IBM.

For most of the twentieth century the punched card remained a standard way of entering data into a computer, and card readers and sorters became an industry of their own. Cards were solid and visible, but they were bulky, easily damaged and awkward to correct, and programmers carried boxes of them between machines. By the 1970s they had largely been replaced by terminals and magnetic storage, and within a decade they had almost vanished from commercial computing.

The loom's contribution to computing can be overstated. It did not calculate, and it had no memory in the modern sense. But it separated a set of instructions from the material being worked on, stored them on a portable medium and allowed the sequence to be altered without rebuilding the machine. Jacquard's apparatus was in that sense the first widely used programmable machine, and the pattern in the cardboard a direct ancestor of the data and programs on which later computers would run.

The reactors that supply most of the world's nuclear electricity are enormous. A single unit typically produces a thousand megawatts or more, enough to supply a large city, and the plant that houses it may take a decade to build and cost tens of billions. The International Atomic Energy Agency classifies anything below 300 megawatts of electrical output as small, and a small modular reactor, or SMR, is an attempt to shrink the unit while changing how it is made. Instead of assembling a unique plant on site, manufacturers build repeated modules in factories and transport them by road, rail or ship for final assembly.

The physics is unchanged. Inside the core, neutrons split atoms of uranium-235, releasing heat and further neutrons in a chain reaction; a coolant carries the heat away and, in most designs, raises steam to drive a turbine. What distinguishes SMRs is scale and architecture. Most current proposals use ordinary pressurised water, in which the coolant is kept liquid at high pressure. Others use gas at high temperature, molten salt or liquid metal, and a few are fast reactors, which use neutrons travelling at high speed rather than slowed down by a moderator.

Smaller cores change the safety problem in useful ways. Heat generated in a reactor after it shuts down, from the decay of radioactive fragments, must be carried away or the fuel will overheat. Large plants rely on pumps driven by emergency generators. Several SMR designs instead rely on passive systems, in which coolant circulates by natural convection and water drains by gravity, needing no power at all. Because a small core has more surface area for its volume, its residual heat escapes more easily, and some designs operate at pressures low enough that escaping coolant would not instantly turn to steam.

Fuel is another area of innovation. Conventional reactors use uranium oxide pellets sealed in metal tubes. Some advanced SMRs use tristructural isotropic particles, known as TRISO, in which a kernel of uranium is wrapped in layers of porous carbon, silicon carbide and dense carbon. These layers are designed to hold fission products inside the particle even at temperatures well above those at which metal cladding would fail. TRISO fuel can be packed into graphite pebbles or into blocks, and its manufacturers argue that it simplifies the containment needed around the core.

Two machines have already been connected to grids. The Akademik Lomonosov, a floating plant moored at Pevek in the Russian Arctic, carries two reactors of 35 megawatts each and began commercial operation in 2020; its heat also supplies a district heating system. In China, a demonstration plant at Shidao Bay in Shandong uses two high-temperature gas-cooled reactors, each producing 250 megawatts of heat, to drive a single turbine of about 210 megawatts, and it began feeding electricity to the grid in 2021.

The economic argument is more contested than the engineering. Optimists point out that a factory can build the same module many times, learning to do it faster and more cheaply, and that modules can be added one at a time as demand grows, so that investors do not have to raise the whole sum at once. Critics reply that a small unit loses the economies of scale that made large plants viable, and that the cost per megawatt is therefore likely to be higher unless production volumes are very large. In 2023 a project to build a group of modules for municipal utilities in the western United States was cancelled after its estimated costs rose sharply.

Regulation is a further obstacle. Each design must satisfy a safety authority, and regulators have had to adapt rules written for large light-water plants. In 2023 the United States Nuclear Regulatory Commission certified the first small modular design, submitted by the American company NuScale, and Britain runs a generic design assessment in which a reactor is examined once and then, with local variations, accepted for use.

Beyond electricity, SMRs are promoted as sources of industrial heat for making steel, cement or hydrogen, and of heat for district heating networks, tasks that at present depend largely on burning fossil fuels. Because the heat is delivered at a few hundred degrees, the reactors can replace boilers directly, and a plant can be sited beside a factory rather than beside a river or a coast, as large reactors must be for cooling.

None of this removes the familiar difficulties of nuclear power: the long-lived waste, the cost of decommissioning and the link, real or perceived, between civil reactors and weapons. SMRs will not solve them. What they may do is make nuclear power available in places and quantities where large plants were never an option, and that is a modest claim rather than a revolutionary one.

Copyright is a bargain between creators and the public. In exchange for a limited period of exclusive control over copying and adaptation, an author accepts that the work will eventually pass into the public domain, where anyone may use it. The first modern statute, passed in England in 1710, gave protection for fourteen years, renewable once. The Berne Convention of 1886, as later revised, went further, making protection automatic in member countries and abolishing any requirement to register a work; its minimum term, the life of the author plus fifty years, has since been extended in most jurisdictions to seventy.

That bargain was designed for a world of physical objects. Printing a book required a press and capital; copying a recording required equipment that few owned. Digital technology dissolved the constraint. A file can be duplicated perfectly, at almost no cost, an unlimited number of times, and sent around the world in seconds. The first mass confrontation came with music-sharing services at the end of the 1990s, and it exposed a mismatch that legislation has never fully resolved: the law regulates copying, but the ease of copying is now a property of the hardware.

Legislators responded by tightening the rules rather than rethinking them. In the United States, a statute of 1998 created a system of notice and takedown, under which a platform escapes liability for infringing material uploaded by its users provided that it removes the material quickly once notified. The European Union went further in 2019, requiring large platforms to make best efforts to obtain authorisation for what their users upload. Both approaches place the burden of enforcement on intermediaries, and both have drawn criticism from opposite directions: from rights holders who say the process is too slow, and from users who say lawful material is removed.

The case for strong protection is not merely the complaint of established industries. A novelist, a photographer or a session musician earns a living from licences, and a market in which copies are free undermines the ability to sell them. Nor is the damage only economic. A writer whose text is reproduced without attribution loses control of its meaning, and an illustrator whose style is imitated at scale by a machine has little recourse in a system built around copying particular works. These are not trivial losses, and they are poorly captured by statistics about sales.

Yet the case for enforcement has been weakened by the way it has been pursued. Global revenue from recorded music roughly halved between 1999 and the mid-2010s, and the decline was widely blamed on piracy. But the industry recovered strongly once legal streaming offered a convenient alternative, which suggests that the problem was as much one of access as of law. Meanwhile, lawsuits against individual file-sharers proved expensive, slow and deeply unpopular, and did little to change behaviour. Suing customers is, in any case, a strange business model for an industry that needs customers to return.

The greater danger, in my view, is over-blocking. Automated systems that compare uploads against a database of protected works cannot tell quotation from theft, parody from plagiarism, or a classroom extract from a pirated copy. They err on the side of removal, because the penalty for failing to remove is a lawsuit. Research, journalism, criticism and education depend on the freedom to quote, and a system that makes quoting risky will quietly reduce the amount of quoting that occurs.

There is also the question of duration. The original fourteen years has become the life of the author plus seventy, a period that no serious economic analysis supports. A term of that length cannot plausibly be an incentive to create, since no author can be motivated by rewards arriving decades after death; it functions instead as a subsidy to estates and corporations. Shortening the term, or requiring renewal after a fixed period, would return a great deal of material to the public domain at little cost.

The alternative to stricter enforcement is not necessarily weaker law. Collective licensing, in which a body collects fees from platforms and distributes them to creators according to how often their work is used, already operates for radio and for many online services. It is imperfect, and it rewards popularity, but it pays creators without censoring users.

Reform should therefore proceed on three fronts. First, safe harbours and platform obligations should be conditional on transparency: companies should publish how many items they remove, at whose request, and how many removals are successfully appealed. Second, the exceptions that allow quotation, criticism, teaching and research should be mandatory and immune to contract, so that they cannot be signed away in a click-through licence. Third, the term of protection should be reduced.

None of this is an argument for abandoning copyright. A society that allows authors to be paid for their work will produce more of it than one that does not, and the digital age has made authorship more accessible, not less. What it has made untenable is the assumption that enforcement by removal is the only tool available. Copyright was always a balance; the task now is to restore it.