The trouble with men

IT MAY be a man's world now, but "Tomorrow belongs to women" (to borrow the title of a book by Jack Lang, a French politician and sex symbol). No wonder.
Apart from being more violent, more prone to disease, more likely to succumb to drugs, bad diet or suicide—more socially undesirable from almost every point of view, in fact—men, it seems, are also slightly more stupid than women. At any rate, boys are doing less well than girls at school. And since, with each passing year, more brainy word-processing consultants and nursery-school teachers get new jobs while more brawny coal miners and machine-tool operators lose theirs, it seems inevitable that women with their graduate certificates and mothering instincts will soon be doing ever more of the world's work, while men lag further behind. A woman's work is never done; a man is drunk from sun to sun.
Well, does that matter? Biologists might answer: not much. To them, men are useful largely for one thing: supplying genetic products to mothers. Providing half a baby's genes stirs up the gene pool and outwits the bacteria and viruses that prey on the species. Also from a biologist's point of view, a lot of aggressive male behaviour is so much genetic advertising. Having men lock antlers, brag about football and indulge in dangerous virility rituals enables women, in some mysterious way, to pick the best genes to hand on.
But nature's methods seem extremely crude. Why not be a tad more scientific? The next generation does not need the current crop of men to be carrying around their sperm all the time. A clean, well-run sperm bank, regularly topped up, would be just as good—and would dispense with men's unfortunate social side-effects. Sperm banks could provide a wide range of gene services, offering, say, high intelligence, predeliction to be a surgeon, blue eyes, long legs and so forth. In America, they already provide a splendid array of choices (and offer insurance in case sperm counts fall even further).
Meanwhile, in terms of cultural evolution, men may well have done their job: they have pretty much set up modern civilisations and technologies; they may not be needed to keep them going. Knowledge-based societies, with their stress on brain not brawn, maybe safer in women's hands.

Testosterone and its antidotes
Back to real life. Men have not been humanely phased out. They still have more jobs than women and, on average, earn higher salaries. In the middle and upper income brackets, men are adjusting—albeit slowly—to women's growing economic power. But down at the bottom of the ladder, where men are men and women change the nappies (but also have the jobs), there are troubles of an entirely different order.
Here whole communities are caught in a deadly vice. As the article on pages 23-32 argues, in areas which used to depend on heavy industry, changes in the nature of work have laid waste the traditional sources of unskilled male employment. Men who used to work in the collieries or shipyards have proved unable (or, which is almost the same thing, unwilling) to do the "women's work" springing up in the industrial ruins. As their jobs have declined, so have their prospects of marriage, for who wants to link their lot with a jobless deadbeat? And as work and marriage have declined together, so everyone has suffered, for these two, since time immemorial, have been the twin responsibilities that have persuaded men to stay with women and children, obey the law and behave as social animals. For women, work and family are often competing spheres; for men, they are linked. When the link is broken, some men, in some places, become loose molecules: uneducated, unskilled, unmarried and unemployed.
Perhaps this will change. Perhaps men will begin to compete more vigorously for "women's work". Perhaps they will find jobs, and, having got them, marry and look after children again. But the evidence so far is against it. At best, men will change their ways reluctantly and more slowly than the quickening pace of economic change. Meanwhile, the blue-collar problem will grow.
Men learn social behaviour through work and marriage, rather than grasp it by instinct. Nothing can be done to change that. Nothing should be done to turn the clock back-by, for example, discriminating against women at work or keeping open economically pointless factories (which would amount to the same thing, since that would penalise the new jobs women do). As the survey after page 64 argues, the engines of economic change—information technology and globalisation—are, contrary to rumour, benefiting the vast majority of humanity: not just women in the West but millions of men and women everywhere.
So such remedies as are possible must concentrate on the places where the social problems are worst: the inner cities. One reason that boys are falling behind girls at urban primary and secondary schools is that too few male teachers are around for boys to look up to and model their behaviour upon. Attracting more male teachers would help close the gap in educational attainment and improve boys'job prospects. More can and should be done to improve the skills of young men: it is especially worth concentrating on their schooling and vocational education, even though men often pooh-pooh such things. And a lot can be done to alleviate the social evils of areas where employment and marriage rates are lowest. While the source of the trouble may be unemployment, much of the damage that flows from this is connected to the easy availability of guns and drugs. Here, more efficient restrictions are needed, which means, for drugs, decriminalisa-tion and control.
Such a list may seem a piecemeal and partial response to "the trouble with men". Granted, it is only a start. But with all great social changes, there comes a stage, before comprehensive policies can be agreed upon, when what is needed is to recognise that a problem exists and accept that it cannot be left to fester. With men in the ghettos, that stage has arrived.

-----------------

MEN
Tomorrow's second sex

The signs are everywhere in America and Europe: more women at work; girls doing better in school; debate about "feminisation" in America's politics; its "million-man march" last year. This article summarises the evidence of a growing social problem: uneducated, unmarried, unemployed men

THESE four pages nail the following arguments to the door of debate:
•  that boys are doing worse than girls at every age in school, except university where girls are narrowing the gap;
•  that women dominate the jobs that are growing, while men (especially those with the least education) are trapped in jobs that are declining;
•  that, for some reason, men are not even trying to do "women's work";
•  that there is a loose connection between work and marriage: joblessness reduces the attractiveness of men as marriage partners;
•  that men do not necessarily adopt "social behaviour" (obeying the law; looking after women and children) if left to themselves; rather, they seem to learn it through some combination of work and marriage (this is a matter of anthropological observation rather than statistical proof);
•  and hence, putting these claims together, that men pose a growing problem. They are failing at school, at work and in families. Their failure shows up in crime and unemployment figures. The problem seems to be related in some way to male behaviour and instincts. It is more than merely a matter of economic adjustment. And (considering the growth in "knowledge-based" employment) it is likely to get worse.

The problem is already far worse in some areas than others. Over the past 30 years, professional men have been less badly affected by economic change than their unskilled brethren. And (to the limited extent they want to) some have added so-called "New Man" attitudes to their traditional breadwinning role. They have adjusted reasonably well to social and economic change. But unskilled men have lost on both counts. Traditional family values-husband winning the bread, wife watching the bairns—tend to be strongest (at least in theory) at the poorer end of the labour market. But American working-class men are increasingly unable (or unwilling) to support families; in Europe, high unemployment has fallen on such men disproportionately. And because providing for a family has been central to men's social role, finding a substitute for steady work will be an immensely hard—conceivably an impossible—task.

Trouble in class
The trouble with men appears early: at school. Though men take up half or more university places in most countries (America is an exception), at primary and secondary school girls are increasingly outperforming boys. In England and Wales, for example, girls score higher than boys in tests conducted at seven, nine, eleven and— which is less often realised—at five. In America, boys are much more likely than girls to be held back a grade and twice as likely to drop out of high school.
Both the reasons for this discrepancy and its true extent are hotly debated. In some subjects at some ages boys still do better than girls (for example, mathematics at 16). Traditionally, boys have done less well than girls before puberty but used to catch up afterwards. What is new now is that boys are no longer catching up. English and Welsh 16-year-olds take a series of tests known as gcses. A standard measurement is the percentage of children who achieve grades a, b, or c in five or more subjects; 48.1% of girls achieve this, compared with 39% of boys. In some of Britain's poorer areas, the disparity is greater. In Hackney, a poor part of east London, for example, a mere 14.9% of boys reached this standard, compared with 30.2% of girls.
The pattern is repeated all over Europe. In 1995, in the European Union, 124 girls got general leaving certificates to every 100 boys. The boys' narrow lead in vocational certificates—they took 5% more—does not close the gap. Girls also tend to stay in school longer: Austria and Switzerland apart, in every West-European country, more girls than boys stay on in education beyond school-leaving age (though the boys who do stay are slightly more likely to go to university, taking 51% of places).

Trouble at work
Because jobs are increasingly "knowledge-based", this disparity in educational attainment is bound to be reflected in employment once today's schoolchildren become adults. This does not necessarily mean that girls have better job prospects than boys; other factors, including sex discrimination at work, may intervene. But it does mean that girls are improving their job prospects relative to boys. Moreover, the job market is already moving the girls' way.
Between 1980 and 1992, women accounted for three-fifths of the increase in the American workforce and two-thirds of the increase in the European one. Between 1990 and 1993, in ten of the then 12 eu members, women's share of unemployment fell.
But the problem for men is not just that women are taking more jobs; it is that a significant proportion of men are dropping out of the job market altogether as women enter it. In the 1960s, almost all men worked and less than half of women. Not so now. The percentage of working-age men in the eu outside the labour force rose from just 8% in 1968 to 22% in 1993. For women, the trend was reversed, falling from 58% to 44% over the same period.
In America, the pattern is slightly different: while women's labour-force participation has risen from 43% in 1970 to about 60% now, men's has dropped relatively little from 80% to 75% (though there is an important exception: male high-school drop-outs—those completing fewer than 12 years of school; in 1970,86% were either working or looking for work; by 1993, only 72% were). If its employment trends continue, America will be employing nearly as many women as men by 2005.
Overall, then, the picture in the West is as follows: the labour market is increasingly friendly to women (though men still make more money and are more likely to be in work); but there are growing numbers of men outside the labour market in a way that women have been accustomed to but men are not.
The future for men looks bleaker, even if you disregard what is going on in schools. Western occupational surveys show that for the foreseeable future new job-growth will be in work typically done by women. America's Bureau of Labour, for example, forecasts that the five fastest-growing kinds of work between now and 2002 will be residential care, computer and data processing, health services, child care, and business services. Women dominate all these activities: their share of employment in them is respectively: 79%, 68%, 70%, 70% and 51%. In contrast, the five sectors declining fastest will be footwear, ammunition-making, shipbuilding, leather-working and photographic supplies. They are man's work. Men account for at least two-thirds of the workforce in all the categories.
There are numerous explanations for female success in growing businesses: women tend to be better educated; they stay in jobs longer (especially women with children); low-paid jobs are growing quickly and women are readier to accept them than men, who still see themselves as a family's breadwinner; women tend to have the social skills needed for jobs in services (though whether because of nature or nurture is disputed).
For men, the obvious response to such economic shifts would be to move into the bits of the economy which are expanding. But they are not doing this. Social theorists may like to claim that concepts like "men's work" or "women's work" are outmoded. The choices people make in the labour force tell a different story. Even in Nordic countries, which have made sexual neutrality a principle of national policy, sex segregation is the norm. Local government, state-run day-care centres, schools and social services are run by women. Men weld cars and take out the rubbish. The pattern is unchanging. In America's "administrative services" (ie, office work) men accounted for only 19.8% of the workforce in 1985 and 20.5% in 1995.
Why should this be? Part of the explanation no doubt lies in the advantages that women have in the workforce—especially their willingness to accept lower-paid jobs. But there seems to be more to it than that. Men continue to spurn even well-paid work that is dominated by women. Less than 5% of America's registered nurses, for example, are men, though the average starting salary of a registered nurse is a comfortable $30,000-35,000. Women account for 96% of America's licensed practical nurses, a responsible but not especially highly-skilled job that pays a full-time worker about $23,000.
The picture is similar in Europe. In Britain, the proportion of men in nursing, 11.6%, has budged little since 1984, when it was 10.2%. As the eu noted in a report on "Occupational Segregation of Women and Men in the European Community", male manual workers are "willing to undertake low-paid and low-skilled jobs provided they are not feminised." In Spain, the share of male office clerks has dropped by a third since 1980—even though this has been a fast-growing field at a time of high unemployment; women, meanwhile, have withdrawn from the textile and footwear industries. In these areas, job separation seems to have increased.

Blue-collar blues
So women are catching up with men for economic reasons ("women's jobs" are growing faster than men's) and social ones (men won't do "women's work"). Both reasons hit unskilled and ill-educated men disproportionately hard.
Jobs that require some tertiary education or training are growing faster than those that require no qualifications. In America's ten largest cities, the number of jobs requiring less than a high-school education has fallen by half since 1970; two-thirds of new jobs created in America since 1989 have been professional and managerial. Germany's Ministry of Labour estimates that by 2010, only 10% of German jobs will be appropriate for unskilled workers. In 1976, the proportion was 35%. In 1970, there were more blue-collar workers than white-collar ones in more than half the oecd countries; by 1990, that was true only in Spain.
In principle, unskilled men could accept these changes and kiss their wives goodbye  on  the  doorstep  as  the  little woman goes off to work at the nursing home. In reality, that is not happening. Despite huge social change during the past 30 years, traditional sexual attitudes retain a stubborn hold. A survey for the eu found that more than two-thirds of Europeans (ranging from 85% in Germany to 60% in Denmark) thought it better for the mother of a young child to stay at home than the father. Mothers, said this survey, should take care of nappies, clothes and food; fathers are for money, sport and punishment.

Trouble at home
Among the poor, this combination of traditional sexual attitudes and male unemployment has been deadly to two groups: men in general in high unemployment areas and, especially, young unemployed men there. The reason is that the combination has set off a spiral of harmful and sometimes uncontrollable consequences which is tearing the web that ties together work, family and law-abiding behaviour.
Consider for a moment a neighbourhood in which most working-age women are not in paid jobs. This may conjure up a picture of tidy homes, children at play and gossip. Now think of a neighbourhood in which most men are jobless. The picture is more sinister. Areas of male idleness are considered, and often are, places of deterioration, disorder and danger. Non-working women are mothers; non-working men, a blight.
Men tend to commit most crimes. In America, they commit 81% of all crime and 87% of violent crime. Adolescent boys are the most volatile and violent of all. Those under 24 are responsible for half of America's violent crime; those under 18 commit a quarter. The figures for most western countries are comparable.
Now ask yourself what restrains such behaviour? The short answer is: a two-parent home. Without belabouring the complexity of family policies, two-parent families are demonstrably better at raising trouble-free children than one-parent ones. Fatherless boys commit more crimes than those with father at home; a study of repeat juvenile offenders by the Los Angeles Probation Department found that they were much more likely to come from one-parent backgrounds than either the average child or than juvenile criminals who offended once only.
Having a man in the house (preferably the biological father) is, it seems, more important than any other single factor. William Galston and Elaine Kamarck, two social scientists who worked in the Clinton administration, argue that the connection between crime and having a father at home "is so strong that [it] erases the relationship between race and crime and between low income and crime." That is why it is a worry that, in America in 1991, just 50.8% of children lived in traditional nuclear families (families where both parents were present and the children were the biological offspring of both parents). Among Hispanics, the figure was 38%; among blacks, 27%.
But family is not the end of the matter. Work also plays a part, both in its own right and as a means of keeping men tied to families. In 1949, Margaret Mead, an eminent anthropologist, argued that in every known human society, everywhere in the world, the young male learns that when he grows up, one of the things which he must do in order to be a full member of society is to provide food for some female and her young ... Every known   human   society rests firmly on the learned nurturing behaviour of men. When men find it impossible to provide, they also seem to find it difficult to learn the nurturing bits. They may retreat into fundamentalist masculinity—the world of gangs which provide for their members a kind of rule-based behaviour that boys do not get elsewhere. For everyone else (and, in the long run, for boys too), the effects of failing to learn nurturing are universally bad.
For an extreme example of this dynamic, take the studies by William Julius Wilson of mass male joblessness in American inner cities*. Here, for the first time in the West, most men are not working and women are the breadwinners (partly because they are working more than ever and partly because welfare cheques go to them). Mr Wilson argues that joblessness, especially among young men, not poverty is a prime cause of the disintegration of inner cities: "high rates of neighbourhood poverty are less likely to trigger problems of social organisation if the residents are working." Mass unemployment, he claims, destroys the institutions that enforce social behaviour—small firms, clubs, informal networks and, above all, the family.
Mr Wilson demonstrates that, for men, employment is strongly linked to marriage and fatherhood; for a woman, children and work are separate (often competing) are much less likely to form one; their attractiveness as a marriage partner sags. Among the inner-city blacks in Chicago whom he studied, almost 60% of those aged 18-44 have never been married and marriage rates are even lower among jobless black fathers than among employed black fathers. And this is true in America as a whole. Black men born in the early 1940s, who came of age during an era of full employment, were more than twice as likely to marry as those born since the late 1950s, who joined the workforce at a time when blue-collar jobs were falling.
So, in the language of social science, "The uncertainties in the labour market are carrying over into uncertainties in the marriage market," as John Ermisch of Britain's Economic and Social Research Centre puts it. Work, it seems, helps determine other aspects of men's lives.
Obviously, the impact upon a neighbourhood of large numbers of jobless single men is influenced by other factors, especially the prevalence of sophisticated deadly weapons. Europe has nothing like Chicago's South Side, the so-called "black belt" that is Mr Wilson's particular area of study. Compared with the South Side, drug use is lower in European cities; gun ownership is much lower; crime is lower; fear of crime is lower. The living standards of the poorest are generally higher. Schools do not have metal detectors.
All the same, if male joblessness is the crucial factor in neighbourhood decline (as far problems, though they may be~bottled up. Not only is average unemployment more than twice as high as in America, but a jobless European stays that way longer. Of Europe's 19m unemployed, more than half have been out of work for a year or more, compared with about 10% of Americans. And young people make up a disproportionate share of Europe's unemployed (except in Germany) and take even longer to get a job. In France, unemployment among under 25-year-olds is 27%, compared with a national average of 12.5%. More than 60% of European youths who are long-term unemployed have never had work; 40% will take two years or more to find their first job. You do not need a sociology degree to worry about the effects of so many young men with nothing to do.

The Japanese solution? No, thanks
There is one rich country that does not have these problems; where just 1% of children are born to single mothers; where crime is low; where marriage rates are relatively high—and where the labour market is rigged in favour of men. This is Japan. In any recession, the "office flowers" are made redundant first. Women are expected to give up jobs on marriage. In the professions, there is not so much a glass ceiling as a concrete one: hard to miss, painful to hit.
The trouble is that the West is unlikely to copy such a system, which is showing cracks in Japan, too. Just because men—or some of them—are struggling in work and at home, women are not about to stride back to the past, accepting the kitchen and nursery as their allotted spheres.
That is all to the good. But there has been a loser in women's march to something closer to equality and that is the man in the blue-collar uniform. Many of the gains that the West has made through enhancing the economic position of women will be tarnished if the male labourer is pushed to the margins. Once known as the salt of the earth, at the moment his troubles are making countries lose their savour.

-----------------

The weirdest computer of all

A quantum computer would rely on the surreal behaviour of the very small to work miracles with information. There is new hope that it might someday be more than fantasy

THE computer has become so widespread, its applications so boundless, that it is easy to forget that it is just a machine, constrained like everything else by the laws of physics. But a machine is all it is, and a pretty simple one at that. The underlying logic of the programs—the binary choices that control the way a computer counts and calculates—would work just as well (if a lot more slowly) in a device that shuttled ball-bearings instead of electrons.
What, though, if a computer could be built around a different kind of physics, the quantum physics that governs the weird and uncertain behaviour of atoms and sub-atomic particles? A group of physicists and mathematicians that is working on this question believes that quantum computing could change computing just as thoroughly as quantum physics changed the Newtonian certainties of classical physics earlier in the 20th century.
Naturally, building a useful quantum computer would be difficult. Only the very smallest objects behave in a detectably quantum way. This means that the components of such a computer would be very tiny and very delicate. But so great is the theoretical appeal of the machine that America's Defence Advanced Research Projects Agency (darpa) has just created an Institute for Quantum Information and Computing, and given it $5m to investigate the possibilities.

Bit players
What makes quantum computing so attractive? For some jobs, a quantum computer would be able to put even the best of today's supercomputers in the shade. It would do this by employing a trick that ordinary computers cannot truly master—parallel processing. Although supercomputers are fast, they can, like all other digital computers, do only one thing at a time. Even the so-called "parallel" computers now available are just collections of machines that each do only one thing at a time. But a quantum computer would be truly parallel. It would be able to do many different things simultaneously in the same piece of equipment. To see why, remember that every digital computer, no matter how wondrous its applications, consists merely of a lot of switches that are linked together and are either on or off. These switches can be thought of as the digits of numbers, in which case their ons and offs are the ones and zeros of binary counting. This is why, by a fortunate contraction of the words "binary digit", the currenc of data processing is known as the "bit".
The switches inside a computer can also be interpreted as the trues and falses of a form of logic known as Boolean algebra. Indeed, they are organised into "logic gates" that do calculations using this algebra. A digital computer is mainly a machine for altering bits by running them through logic gates. The gate known as and, for example, compares two bits. If both have a value of "one", it creates an output of one; otherwise it creates a zero. With the next tick of its internal clock, the computer then passes this output bit to another gate for further processing. Each tick, then, corresponds to one step of the calculation.
In the quantum world, however, the rules are different. A quantum computer's switches would not have to be either on or off. Like any quantum object which can come in several distinct states, a quantum switch can carry on indefinitely as a combination of all those states. It is not merely stuck halfway between on and off; it is actually on and off simultaneously, as if it existed in two parallel worlds.
The bits in a "quantum computer" would not, therefore, be ones or zeros. They would be quantum combinations of one and zero. Such vacillating pieces of information are known as "qubits" (pronounced as "cubits", though they are usually rather smaller than the ancient Hebrew unit of length).
It is this simultaneous existence in many states that would give quantum computing its power. An ordinary computer having (say) ten bits could exist in only one of 1,024 states (all the ten-digit sequences that can be made from ones and zeros) available to it, at each instant. A quantum computer with ten qubits could exist in all 1,024 states at the same time. It could therefore work on 1,024 calculations at once.
But qubits are fiendishly hard to make. For a quantum switch to maintain its multiple existence depends on one big proviso. Nothing can happen that might disturb it. No light must shine on it, no stray air molecules may bump into it, no inquiring probe can intrude on it. If it is disturbed, it stops vacillating and chooses a definite state to be in. This choice depends on chance and on the way its multiple personalities have interacted until then. If it happens prematurely, the computation will be ruined.
This proviso has made technical progress agonisingly slow. Paul Benioff, at Ar-gonne National Laboratory, in Illinois, first applied quantum theory to computers in 1981. David Deutsch, at Oxford University, proposed the possibility of quantum parallel computers in 1985. Yet both had to wait until last year for the first two qubits to come on-line.
One group of quantum mechanics who created them was led by Chris Monroe, who works at the National Institute of Standards and Technology, in Boulder, Colorado. Dr Monroe's group trapped a single ion (an atom with missing electrons) of beryllium, by walling it in with electric and magnetic fields and cooling it to within an ace of absolute zero (-273°C). Such an ion has two energy levels—providing the first qubit. And it can vibrate in two ways within the trap—the second qubit. These qubits are linked together via a pulse of laser light, which can switch the ion from one energy level to the other. Whether it does so or not depends on which of the two vibrational patterns the ion is in.
The result is a logic gate called xor. This has two input bits, and reverses the value of the second if the first is "one". That is an important step. Though using many types of gate makes designing ordinary computers easier, it is possible to make do with just three: and, not and copy. A quantum computer is even simpler. It requires only two sorts of gate, and xor is one of them. (The other would be able to change a zero into an arbitrary combination of zero and one. This has no equivalent in conventional computing.)
Qubits can also be made out of light, because light can be polarised in two perpendicular directions. Jeff Kimble, a physicist at the California Institute of Technology (Caltech), in Pasadena, and the head of the darpa quantum-computing institute, is working towards a light-based device. He and his team have designed a piece of apparatus that allows photons (the particles of which light is composed) to interact while they fly through a stream of caesium atoms. This interaction could in principle form the basis of logic gates like xor.
Trapping ions, though, may ultimately prove to be the best way to make real quantum computers. Such devices would require thousands of ions, vibrating in synchrony, and are certainly still in the realms of science fiction. But ion traps housing up to six ions have already been built during other atomic-physics projects. Groups at the Los Alamos National Laboratory, in New Mexico, and the University of Innsbruck, in Austria, are trying to improve upon them to connect a handful of qubits together.

Qubits v true bits
Why bother? Why should darpa, better known for researching missile defences and commissioning the precursor of the Internet, dabble in quantum theory? Perhaps to make better calculators. There are already three mathematical challenges known in which quantum computers would beat the pants off the ordinary kind.
The first is up darpa's alley: cracking secret codes. The security of most "public-key" coding schemes, in which people can exchange secrets without advance planning, relies on inscrutable large numbers. It is hard to factorise—to find numbers that multiply together to produce—some 200-digit numbers. But it is these factors that unlock the code. Since the programs for finding them are excruciatingly tedious, even for a fleet of supercomputers, such codes are pretty safe.
A quantum computer, on the other hand, would make short work of this task. In 1994, Peter Shor, a mathematician at at&t Labs, in Murray Hill, New Jersey, designed a quantum circuit that could achieve it in many fewer steps. His quantum software arranges the multiple personalities of the qubits in a way that enables them to attack the problem separately, and then conspire together so that when they are, in the end, forced to behave like normal bits, they will be likely to produce the correct answer.
Like much of quantum theory, this can be properly explained only in mathematics, not words. But a sense of what it entails can be gained by recognising that the equations of quantum theory are not unlike those of wave motion, as Seth Lloyd, of the Massachusetts Institute of Technology, has suggested. Each parallel path of a qubit is analogous to a sound wave of a particular pitch. When different tones play together, they interfere, reinforcing or cancelling one another depending on their strengths and timing. The parallel lives of qubits do the same. And the strongest tones are the ones likeliest to be chosen by the qubits when they are measured.
In this analogy, an ordinary computation can be thought of as a melody of single notes plinked out on a piano. A quantum computation, with its complex array of chords and rhythms, sounds more like the Berlin Philharmonic. Dr Shor's trick was to arrange the concert hall so that the music was completely dominated by the proper overtones and beats—the rich timbre of the French horns, say—that correspond to the correct solution.
Lov Grover, a computer scientist at Bell Labs (which recently split from at&t Labs), has written quantum symphonies to solve two other problems. One, published in May, is a way to search unsorted databases. If you were to sift through 10,000 pieces of paper scattered about your desk in search of one important memo, you would have to be prepared to scan 5,000 of them to have an even chance of finding it. By assigning each parallel computation to pursue a different choice, a quantum computer could do the same in only 40 runs.
Dr Grover's other discovery, announced this month, is that quantum computers would be talented statisticians. They would excel at finding single numbers which depend collectively on lots of data: the median age of a large population, for example. If a quantum computer and an ordinary computer with the same internal clock rate raced to do it, the quantum computer would win hands down.

Back to the future
There is, however, a snag. Qubits, not being true bits, are not truly digital. One of the big advantages of recording and processing information digitally is that it is difficult to make a mistake (electronic "ones" are not that easily turned into "zeros"), and surprisingly easy to design ways to recognise and correct any mistake that is made.
This ease of correction was shown in the 1940s, at a time when the first digital computers were being constructed. Claude Shannon, who also worked at at&t, set the ball rolling by investigating ways of encoding bits of information so that they would resist errors during transmission down telephone lines. In the 1950s, John von Neumann, one of the mathematicians who helped to develop the theory of these early computers, went on to show how a reliable computer could be built out of unreliable parts, by using Shannon's ideas and combining them with extra (so-called redundant) circuitry.
The simplest error-correcting code is triple-redundant. Each bit is copied into two more bits: o becomes 000, and 1 becomes 111. If one bit accidentally flips, the other two indicate this and can be used to fix it. But Rolf Landauer, a physicist at the ibm Watson Research Centre, in Yorktown Heights, New York, has insisted that this cannot work with qubits, which cannot be measured, let alone copied. Yet error-correction would be vastly more important in a quantum computer than an ordinary computer, since the disturbance of even one qubit could ruin the coherence of the whole computation. If quantum computing is to become more than a dream, it needs a new way to correct errors as well as new hardware.
It was therefore a welcome surprise when Dr Shor announced a year ago that he had invented a quantum error-correcting code. Instead of three qubits, Dr Shor's encoding formula required nine. A few months later other researchers at ibm, Los Alamos and Oxford came up with codes that work with five of them.
The idea begins with the qubit that needs to be protected from errors, and four other qubits, set to zero. A sequence of logic gates changes and tangles all five bits into a larger quantum combination. Just as one qubit is a mix of two different states, two qubits are a mix of four, three are a mix of eight, and so on. Five entangled qubits span 32 possibilities—each of the five-digit sequences that can be made of ones and zeros.
Five qubits are enough, though, to preserve the quantum ambiguity of what is going on, even if one of them goes bad. One qubit being unmasked by, for example, a collision with a stray air molecule, does not reveal enough information for an observer to work anything out about the qubit originally used to make the mixture. In these circumstances, the calculation can be corrected, and proceed unhindered. The original mixing, in other words, creates quantum cannon fodder  for  such  interactions  as would otherwise spoil a combination. Some losses become tolerable without the game being given away.
There are still problems to be overcome, of course. It is necessary to design logic gates that will work on entangled bundles of five qubits at a time. And errors can still creep in during the encoding process itself. Dr Shor's latest achievements, to be announced at the Symposium on the Foundations of Computer Science, being held in Burlington, Vermont, in October, address both of these concerns. First, he has designed logic gates for protected qubits by combining lots of rwo-qubit logic gates. Second, he has shown, theoretically, that the improvements in components needed to make quantum computing reliable should not, once such components actually exist, be too onerous.
His second result is analogous to the ideas in one of von Neumann's achievements, the "Synthesis of Reliable Organisms from Unreliable Parts". Early sceptics of computing thought that allowing computers to work, say, ten times longer on a problem would require components that were ten times as reliable. Von Neumann showed that, with the shrewd use of error-correcting codes and redundant logic gates, the components did not need to be any more reliable. All that was necessary was that the speed of the calculation be slowed down to allow for error-correction. Dr Shor's result is not quite as powerful as this. He has shown that to allow a quantum computation to proceed for ten times as long, the reliability of components does have to be improved a little. But the improvement required is a small fixed increment, not a factor often.
That quantum error-correction is possible at all has led to a new wave of optimism about quantum computing. The circuitry for error-correction is too complex to be built in a laboratory anytime soon. But the group from Innsbruck, which is led by Peter Zoller and Rainer Blatt, has just published another partial error-correcting code based on the principle of quantum cannon fodder. This one is still simpler than Dr Shor's and should be particularly effective in correcting the sort of errors that might be expected in an ion-trap quantum computer.

Three times five, silly
The Innsbruck group hopes to try out its scheme with a three-bit computer by 1998. Meanwhile, in Los Alamos, Richard Hughes and his team are planning to implement Dr Shor's algorithm for factorisation. With luck they will be able to factorise the number 15 within five years.
That this is considered ambitious hints at the difficulty of this latest merger of physics and computing. Indeed, many physicists doubt whether quantum computing will ever amount to anything useful. Dr Landauer is one of them. His work on the ordinary (not quantum) physics of information led to a new appreciation of information's inextricable relationship with physics. But he does not expect ibm to be mass-producing quantum computers anytime soon.
In this, he is assuredly correct. And even among the quantum computer's enthusiasts, most researchers are motivated less by a desire to produce a practical quantum computer than by a wish to learn how to wield more control over quantum effects themselves. Until now, nobody has been able to direct quantum states with the flexibility or precision required for quantum computers. Doing so should help to illuminate some of quantum theory's more curious features.
In  addition, if quantum  computing really could be made to work, physicists would have a powerful new analytical tool. One use for it would be to explore the many ramifications of quantum theory (such as the "Hubbard model" of how electrons hop around in a crystal, or theories of fundamental particles) that are thought to be correct but cannot be tested precisely. The sort of equations needed to check these ideas are too much for even the most nimble non-quantum supercomputer to handle.
In 1982, the late Richard Feyn-man, who was also at Caltech, suggested that a quantum computer would help to solve these sorts of problems. To simulate an ordinary object, only a few numbers are needed. For a quantum object every possible state must be tracked— an enormous set of numbers. How better to keep up with the calculation than by using something that is itself in a multitude of states?
Whether such uses will come to fruition remains to be seen. But, as Dr Landauer points out, in one sense quantum computing has already been useful. It has provided a salutary reminder that computing does not take place in an abstract mathematical world. It uses earth, air, fire and water—or single atoms of those things. It is subject to all the limitations, and all the possibilities, written in the laws of physics. Some mathematicians and computer scientists spend their days searching for the quickest possible algorithm to solve particular problems. To their surprise they must now learn quantum theory, in order to cover all the possibilities allowed by nature. After a long absence, they are being forced to return from the virtual reality of information science and deal with the real world, in all its indeterminate queerness.

----------