IELTS Academic Reading Practice Test 139

The personal digital assistant, or PDA, was a small battery-powered device that held a calendar, an address book and a list of tasks, and that could be linked to a desktop computer. The name was coined in 1992 by John Sculley, then the head of Apple, at a trade fair, to describe a category of product that barely existed. A PDA was meant to be carried everywhere and consulted many times a day. It was not a telephone, and in its early years it did not connect to the internet.

Handheld computing began with the pocket calculator. By the late 1970s, manufacturers were selling calculators that could store a few numbers in memory, and a small industry grew up around 'electronic organisers': flat devices with a keypad, a one-line display and a database of names and telephone numbers. They were cheap and limited. The machine that changed the category came from Britain.

Psion, a British company founded in 1980 by David Potter, launched the Organiser in 1984. It was a pocket computer rather than a calculator. It had a small keyboard, a one-line liquid-crystal display and a slot for removable memory cartridges, and it ran on ordinary batteries. The Organiser was bought by scientists recording field data, by sales representatives and by anyone who needed a computer that fitted in a coat pocket. A later model, the Organiser II of 1986, added a two-line display, a proper database and a language called OPL in which owners could write their own programs, and it sold in much larger numbers.

Apple's contribution was more ambitious and less successful. Sculley wanted a device that would recognise handwriting, so that its owner could write on the screen with a stylus as if on paper, and he made the project a priority. The Newton MessagePad, released in August 1993, was the result. It had a large screen, a stylus, an infrared port for exchanging data with other machines and software for notes, names, dates and simple drawings. It cost about 700 dollars.

Handwriting recognition attracted most of the criticism. The software compared each stroke with models of how letters are formed, and it was designed to learn the habits of its owner over time. In practice it made frequent errors, particularly with unfamiliar writers, and the errors became a joke in a widely syndicated American comic strip. Reviewers also complained that the device was too large for a pocket and that its price was high for what it did. Early sales were modest.

Later versions were smaller and much more accurate, and a loyal group of users grew attached to them. But Apple was losing money, and when Steve Jobs returned to the company in 1997 he closed the project down. The last Newton was discontinued in 1998, by which time the cheaper Palm devices had been on sale for two years.

Palm Computing was founded by Jeff Hawkins, an engineer who had spent years trying to build a pocket computer and had decided that the difficulty lay in input. The Palm Pilot, launched in March 1996, abandoned handwriting recognition in favour of Graffiti, a system in which each letter is written as a single simplified stroke. The device was small enough for a shirt pocket, ran for weeks on two ordinary batteries and cost 299 dollars. The first production run sold out, and the Pilot established the layout that almost every later organiser copied: a screen above an area for writing, and four buttons below it.

Palm's other decisive innovation was the cradle. The device sat in a small stand connected to a personal computer, and a button on the stand began a process the company called HotSync, which copied changes in both directions so that the two machines held identical calendars and lists of contacts. The desktop computer held the master copy; the handheld held the version that was carried around. This division of labour made the PDA acceptable to employers, who were willing to buy a device that backed up its data onto a company machine.

The next stage moved the organiser into communication. Research In Motion, a Canadian company based in Waterloo, Ontario, began in 1999 to sell two-way pagers that relayed electronic mail. Messages were pushed to the device as they arrived rather than collected on request, and the small keyboard below the screen allowed rapid typing with two thumbs. The BlackBerry name was used for later models, which added a telephone. Its security made it acceptable to banks and to government offices, and for a decade it was the standard tool of senior executives.

Meanwhile the functions of the PDA were absorbed by the mobile phone. Microsoft's Pocket PC system and the Compaq iPAQ put a colour screen and office software into a handheld, and handset makers added calendars and address books to their phones. The decisive step came in 2007, when Apple released the iPhone, operated by a finger rather than a stylus and supplied with a web browser. Buyers who had carried an organiser and a phone began to carry one device.

Palm released a new operating system in 2009, but the company was sold to Hewlett-Packard in 2010 and the devices were withdrawn. By then the separate organiser had effectively disappeared: the calendar, the address book and the synchronised copy held on a distant server had become features of a telephone rather than of a distinct product. The PDA had not failed; it had been absorbed by something that could also make calls.

Steel is an alloy of iron with a small proportion of carbon, usually less than 2 per cent, and it is the most widely used structural material in the world. It is strong, can be shaped and welded with relative ease, and can be melted down and reused almost indefinitely. Global production of crude steel is now close to two billion tonnes a year. The industry is responsible for roughly 7 per cent of the carbon dioxide that human activity releases into the atmosphere, and most of that comes from the way iron is extracted from its ore.

In a traditional steelworks, iron is released from its ore in a blast furnace, a tower lined with heat-resistant bricks and fed from the top. The raw materials are iron ore, which is largely iron oxide, coke made from coking coal, and limestone. Blasts of hot air are blown in near the bottom, and the coke burns fiercely, producing carbon monoxide. This gas rises through the descending ore and strips oxygen from it, leaving molten iron that collects at the base of the furnace. The limestone combines with sandy impurities to form a slag, which floats on the iron and is drawn off separately. The metal tapped from a blast furnace contains about 4 per cent carbon, which makes it brittle, so oxygen is then blown through it in a second vessel to burn off the surplus.

The difficulty is that carbon plays two distinct roles in this process. It provides the heat, which in principle could come from electricity instead, and it also acts as the reducing agent that removes oxygen from the ore. The second role is the awkward one. Every tonne of iron made in a blast furnace inevitably produces carbon dioxide, because the carbon combines with the oxygen it has taken from the ore. Capture and storage of the gas can reduce the emissions, but it cannot eliminate the chemistry.

Hydrogen offers a way out because it can perform the same task. If hydrogen is passed over iron ore at a temperature of several hundred degrees, it combines with the oxygen to form water vapour and leaves the iron behind as a porous grey solid known as sponge iron. The reaction is well understood and has been used for decades in plants that burn natural gas to make hydrogen; what is new is the idea of supplying the gas from renewable electricity rather than from fossil fuel. The sponge iron is not liquid, so it must then be melted in an electric arc furnace, usually together with scrap, before it can be cast into slabs and rolled into sheet.

Electric arc furnaces already account for a large share of world production: roughly 30 per cent of steel is now made by melting scrap, and that route can run entirely on renewable electricity. Scrap, however, is not available in unlimited quantities, and much of it is contaminated with copper from old wiring and motors, which is difficult to remove and degrades the quality of the steel. If demand grows, and if recycling cannot meet it, some iron made from ore will always be needed. Green steel plans therefore depend on direct reduction rather than on recycling alone.

Several projects are now testing the idea at scale. The best known is HYBRIT, a venture formed in 2016 by the Swedish steelmaker SSAB, the state-owned mining company LKAB and the utility Vattenfall. A pilot plant at Luleå in northern Sweden began test operations in 2020, and in 2021 the partners announced that it had produced sponge iron using hydrogen rather than coke. A second Swedish company, Stegra, formerly H2 Green Steel, is building a plant at Boden that is intended to run on hydrogen from electrolysers. Steelmakers in Germany and Spain are also carrying out trials.

The obstacles are formidable, and most of them are to do with energy and ore. Making hydrogen by splitting water with electricity is itself inefficient: a large share of the electrical energy is lost at each stage, first in the electrolyser, then in compressing or liquefying the gas for storage, then in heating the furnace. A plant producing a million tonnes of steel a year would need several gigawatts of dedicated generating capacity and a supply of exceptionally pure ore, because direct reduction, unlike a blast furnace, cannot cope with ore that contains much silica or alumina. Ore of that grade is mined in only a few places. Even the most optimistic plans therefore assume that green steel will be made where electricity is cheap and plentiful, which in practice means countries with very large supplies of renewable power.

None of this means that green steel is a fantasy. The chemistry works, and the first commercial plants are being built. But the cost of a tonne will be higher than that of conventional steel for years, and someone must pay what the industry calls the green premium. In Europe, the emissions trading system already forces steelmakers to buy permits for the carbon they release, which narrows the gap, and a carbon border adjustment mechanism is intended to prevent cheaper, dirtier imports from undermining European producers. Recycling, better furnace design and carbon capture on existing plants will all contribute. The decisive question is not whether steel can be made without coal, but how quickly the world's furnaces can be converted.

Whenever a technology company becomes large enough to attract political attention, someone reaches for the example of Standard Oil. In 1911 the United States Supreme Court ordered that company to be divided into more than thirty separate firms, and the decision is still cited as proof that size can be confronted. The comparison is tempting, but it obscures more than it explains. Oil was a physical commodity sold by the barrel, and a customer who disliked one supplier could buy from another. The largest digital platforms are not like that. They are markets as well as companies: places where other people's businesses and conversations happen, and where the rules are set by the owner of the place.

The tendency towards concentration is built into the economics of these services. A messaging service becomes more valuable to each user as more of their friends join it; a marketplace attracts sellers because the buyers are there, and buyers because the sellers are. The company that owns the largest network also collects the most data, which it can use to improve its product and to target advertising, and it can spread the fixed cost of its software across a customer base no newcomer can match. Such a market may look competitive on paper, with no legal barrier to entry, yet be extremely difficult to enter in practice.

It would be a mistake, though, to treat size as proof of wrongdoing. These companies grew because they made things that millions of people freely chose to use, and many of their services cost the user nothing at all. Their scale has paid for research that smaller firms could not have afforded, and it has driven down the price of communication, mapping, translation and data storage. Consumers have gained a great deal from the concentration of these industries, and any policy that ignores those gains will not survive contact with the public.

The standard defence of the industry is that the market remains contestable. Incumbents can be swept away, the argument runs, and they often have been: the makers of mobile phones that dominated in 2007 were not dominant five years later. That is true, but it is less reassuring than it appears. The firms at risk in that example were selling hardware, a business in which a better design can win in a single season. The platforms that now control search, social contact and online retail are protected by the habits of their users and by the sheer inconvenience of leaving.

A young company with a genuinely better idea therefore faces a choice. It can try to compete with an incumbent that controls the channel through which most customers are reached, or it can sell itself to that incumbent for a sum its founders would never earn otherwise. Many have chosen the second. Buying a potential rival is usually cheaper than fighting it, and the purchase removes a competitor from the market before it has had the chance to become one. This is the strongest argument for intervention, because it attacks the mechanism by which competition is quietly reduced rather than the size of the winner.

Size also brings a power that has nothing to do with price. Because these platforms decide what appears in a feed and what is removed from it, they exercise an influence over public conversation that was once reserved for newspapers and broadcasters. The companies dislike this responsibility and insist that they are neutral conduits, but their algorithms are editorial choices by another name. Some politicians respond by demanding that the state decide what is permissible instead, which would replace private censorship with a far less accountable version.

Regulation is therefore necessary, but its form matters. The European Union's Digital Markets Act, adopted in 2022, relies not on breaking companies up but on obligations. Designated platforms must offer data portability, so that a user can take an account elsewhere; they must permit other businesses to reach customers on equal terms; and they must not favour their own services in rankings. In my view this approach is more promising than structural separation, not because concentration is harmless but because it goes after the harm itself. If moving to a rival becomes genuinely easy, the incumbent must keep earning its users.

The most useful reforms are therefore modest in appearance and radical in effect. Rules that require data to be portable, that oblige a platform to work with its competitors' software, and that cut the cost of switching from one service to another do not ask a regulator to guess how large a company should be. They simply lower the walls that keep users inside. Interoperability is the key, because a network's hold over its users lasts only while it is closed.

Governments have become dependent too, storing records and buying computing from the same handful of suppliers. A failure in one provider's data centre can interrupt hospitals, banks and government offices on the same afternoon. Competition policy is therefore no longer only about prices; it is about who is able to speak and who is able to be heard.

None of this will satisfy those who think only a break-up can restore competition. I understand the impatience, but the history of antitrust suggests that structural remedies are slow, contested and difficult to reverse once a market has changed shape. The test of policy should be narrower and more practical: can a better product, offered by a stranger, reach its users without the owner of a platform obstructing it? On that test, the largest technology companies have not yet been asked to prove themselves, and it is time they were.