Perhaps the boldest theme to have emerged from the Government this year is the idea that Britain needs more immigrants. Back in July, the Immigration Minister, Barbara Roche, heralded a radical change in policy when she told a conference in Paris: "Throughout the centuries, immigrants have had a very positive impact on the societies they join. We need to find ways to meet legitimate desires to migrate, and be ready to think imaginatively about how migration can meet economic and social needs."

Perhaps the boldest theme to have emerged from the Government this year is the idea that Britain needs more immigrants. Back in July, the Immigration Minister, Barbara Roche, heralded a radical change in policy when she told a conference in Paris: "Throughout the centuries, immigrants have had a very positive impact on the societies they join. We need to find ways to meet legitimate desires to migrate, and be ready to think imaginatively about how migration can meet economic and social needs."

But before we start getting too excited about this softening of hard northern hearts, this throwing-down once again, after 30 years, of the national "welcome" mat, it's worth looking a little harder at what the proposed new policy is going to mean.

It is not just any old migrants that Britain wants, of course, but the best brains the developing world has to offer. There is, for example, a strong demand for the people with IT skills that India produces in large numbers. And there is no doubt that a great many of them would love to come to work in the UK and probably settle here for the rest of their lives.

But the consequence could be to help to turn India's already chronic brain drain into a flood, and to push even further into the future the possibility of the benighted subcontinent fulfilling its potential.

India has an ancient genius for mathematics. Zero was an Indian invention, and so, too, was the Pentium microprocessor. Moreover, high-technology tertiary education has been one of the most heartening - and most surprising - success stories of India's first independent half-century.

The symbols of that success are the six Indian institutes of technology (IITs), which were set up after independence, the first one opening in 1951, in frank imitation of the Massachusetts Institute of Technology.

They were one of the great nation-building initiatives of the Nehru years, intended, like the big dams, central planning and the so-called permit Raj, to help to thrust India out of the ranks of the poor and needy and into the company of the top nations, where she belonged.

None of Nehru's other big ideas has stood the test of time. The IITs, however, have done very well. Nearly half a century since they began taking in students, they remain among the best technical universities in the world.

They have hung on to that reputation despite extraordinary handicaps. The campuses are pleasantly leafy, but they are run down and ill equipped by comparison with even mediocre universities in the UK.

In a hall of residence in one institute, for example, only 30 out of 300 students own their own computers, and none of those computers has internet access. Two old-fashioned telephones serve the entire building. Although their successful graduates have recently begun making donations, they are still impoverished schools: they spend the equivalent of about £1,500 per student per year. Student fees per annum come to less than £600.

Few students have the money to buy textbooks. Instead, they typically share a single book borrowed from the library among half a dozen students, with each given a fixed amount of time in which to digest it.

Students compensate for their lack of resources and the shortcomings of their environment by working ferociously hard. Of the 125,000 candidates who take the entrance exam every year, only 3,000 are accepted - fewer than 3 per cent, which compares with the rate of acceptance at Harvard of 16 per cent. Once admitted, they commonly study until 3am. Teaching is old-fashioned, with frequent tests and merciless grading.

Why do they submit to such a regimen? Nehru would have liked to believe that it was all in the cause of making India great. But the bitter fact is that the typical IIT student is striving to escape from India altogether.

Very many of them succeed: at present, around one-third of IIT graduates end up in the West, with the United States being the most favoured destination. When they get there, many of them flourish. One of the early migrants, Preetinder S Virk, who graduated from IIT Kharagpur in 1962 and is now a professor of chemical engineering at MIT, says, "The best IIT students are always damn good. And in America they flourish. It is an environment oriented to doing good work."

Sweating over their borrowed books and primitive, shared computers in the non-air-conditioned fug of an IIT dormitory in Bombay or Madras, the students have plenty to dream about. Some 30,000 IIT alumni are settled in the US. They include people such as Gururaj Deshpande, founder of IT companies including Cascade Communications, with a personal net worth of about $3bn; Vinod Khosla, co-founder of Sun Microsystems; Rakesh Gangwal, chief executive of US Airways Group; and Suhas Patil, founder of Cirrus Logic Inc.

By one reckoning, fully 10 per cent of companies started in Silicon Valley between 1995 and 1998 were the work of Indians, most of them IIT graduates.

The result of Barbara Roche's promised liberalisation is likely to be that a large number of the IITs' budding geniuses will cast their eyes not to Silicon Valley but to Silicon Glen and Silicon Fen instead. The advantage both for Britain and for the new arrivals is plain.

But where does all that leave India? In the past decades, India has miserably failed to make any sort of a mark in manufacturing, despite her huge domestic market and an ocean of cheap labour. So, in the past few years she has persuaded herself that national salvation lies in IT. Last week, Bill Gates flew into Delhi, holing up at the Maurya Sheraton Hotel throughout his brief stay: he drove practically everything else off the Indian newspapers' front pages. The chief ministers of 10 states, with a combined population of half a billion people, danced attendance on him, eager to catch his eye. After the success of Bangalore and Hyderabad, two southern cities that are home to numerous software development companies, the rest of India has decided that this is the only way the country is going to get rich.

It would be laughable if it were not so sad. India's IT success is remarkable - but only against the backdrop of the lamentable failure of her other industries. IT has been spared (until recently) the dead hand of central government regulation. But its achievements in absolute terms are both modest and fragile.

India's software industry shot to prominence because of the need for large numbers of software engineers to crunch numbers in the run-up to Y2K. India has lots of moderately well-trained engineers who are moderately proficient in English and who will work for next to nothing. It's all low-end stuff. As one Indian commentator put it bluntly last week, "Indians are the software coolies of the world."

Threatening to undermine even this flimsy success is the atrocious Indian infrastructure, which makes the simple act of logging on to the internet in India an endless nightmare. With far better provided countries such as Malaysia eating into India's linguistic advantage, there is no guarantee that the success India currently enjoys will not soon dribble away among the potholes and burnt-out wiring.

None of that is to argue that Barbara Roche should not relax immigration controls. From a selfish point of view, it's an excellent wheeze and probably long overdue. But let's not kid ourselves that we are doing India any favours.