(The original version of an essay published in Prospect, April 2004 )
Pontificators at smart dining tables in London and
Washington D.C. hopelessly confuse cause and effect in
debating outsourcing and the reasons for India’s rise
to prominence as a source of software brainpower. “It
all began with those Indians in Silicon Valley, didn’t
it?” they say. Or, “Of course it’s those Indian
technical colleges, those IITs – what, half a dozen
clones of the Massachusetts Institute of Technology? –
turning out programmers and computer scientists by the
container load.”
The best that can be said of these
attempts at explanation is that they are an advance on
the idée fixe of a year ago, which was that Western
companies were only paying Indians and Indian
companies to write and design software because they
are cheap – because programmers and software engineers
in India can be hired for roughly a fifth of what they
would earn in America.
The actual reasons for India’s soaring stock in
software are complicated, border on the bizarre, and
are rooted in the subcontinent’s earliest intellectual
and pedagogical traditions.
Software is ubiquitous and all-pervasive. It is at the
core of processes and tools in every strategic
industry – from banking and finance to microchips and
defence. The depth of India’s comparative advantage in
software suggests that India poses a far more
interesting challenge to the Western economies than
even China does. China, strong in manufacturing and
evolving into a powerhouse in computer hardware
production, has been almost as unimpressive in
software as Japan – and India the reverse. No
underdeveloped country has ever taken on the developed
world in a craft as sophisticated and important as
software.
Indian software aptitude rests on an unlikely pair of
factors: an emphasis on learning by rote in Indian
schools, and a facility and reverence for abstract
thought. These biases of Indian education are all but
mutually exclusive in the modern West, where a
capacity for abstraction is closely associated with
creativity and stimulating, inspirational learning. In
India, learning by rote is seen by many, if not most
conventional teachers, as essential grounding for
creativity – like Picasso’s mastery of perspective and
anatomy in his youth – and for unbounded invention and
speculation.
An educational tradition that spans learning by heart
and exalting excellence in higher mathematics and
physics as the height of academic accomplishment is
just right for software. It perfectly fits the
mentality of computers. These are, after all, machines
so rigidly conformist as to refuse to send email with
a missing hyphen or full stop in an address. Yet no
product on earth is more abstract, boundlessly complex
and flexible as software, which cannot be seen, heard,
smelled, tasted or touched – is, to borrow from
Vladimir Nabokov’s take on chess, a game invented in
India, “a spectral art.”
India’s software accomplishments to date reflect those
extremes. At one pole, Indian firms overwhelmingly
dominate a world élite of over 120 companies
recognised for producing outstandingly accurate
software – having earned a CMM Level-5 tag, or
software’s equivalent of the Michelin 3-star rating
for restaurants. These establishments – of which
America has less than half the Indian total – are
formally certified to be following an exacting,
tediously detail-obsessed methodology developed at
Carnegie-Mellon University in Pittsburgh for producing
reliable code.
At the other pole, most of the world’s reigning
American technology giants – Microsoft, General
Electric, Texas Instruments, Intel, Oracle and Sun
Microsystems – have established software development
facilities and even R&D laboratories doing fundamental
research on the subcontinent. They see India as a
bargain basement for an unlikely commodity,
world-class technical and scientific brains –
engineers, computer scientists and researchers. A top
electrical engineering graduate from an Indian
Institute of Technology earns about an eighth of the
starting salary of an American counterpart.
The most far-sighted Brahmin sage of circa 1500 B.C. –
roughly when the earliest of the Vedas, Hinduism’s
sacred über-texts, are thought to have been written
down – could not have envisioned any such application
of the teaching conventions born at the same time.
Exactitude was of the essence of the pedagogy of the
Brahmins, fiercely exclusionary scholar-priests,
because their pupils were not merely acolytes but,
effectively, human zip disks – data storage media. The
Vedas were preserved and passed down orally for many
centuries (some Indian scholars claim, thousands of
years) before they became texts. That meant that an
exemplary Brahmin scholar of the time had to be
capable of holding in his head the equivalent of
several books of the Bible and scholarly commentaries
on them and an entire Sanskrit thesaurus.
Preservation aside, exactness in memorization mattered
because, as Nicholas Ostler of the Foundation for
Endangered Languages has explained, Sanskrit mantras
“are Vedic hymns used as incantations,” or as “sacred
formulae whose incessant repetition was held to have
important mystical effects.” A priestly acolyte had to
be capable of not just a word- but phone-perfect (as
in phoneme) recitation of them, with the proper
intonation – because different sounds corresponded to
different spiritual purposes.
The precise, specialised languages we use to program
computers are, like hieratic Sanskrit, deployed to get
absolutely specific results considered vital by their
users – even if they serve largely economic ends, and
mantras religious ones. Many details of computer
languages and their rules – and variations of these
for different contexts – have to be memorized by
computer programmers. The ability to retain details of
large chunks of indigestible information is just as
critical for a programmer as for traditional Brahmin
scholars.
Although Indian children no longer learn by rote to
serve as zip disks, and Sanskrit – when it is taught
in modern Indian schools – has much the same museum
quality as Latin and Greek in British classrooms, rote
learning still holds tremendous sway on the
subcontinent. As newspaper editorials routinely
attest, the most common lament about the state of
Indian education is about the continuing reign of
rote. Children commit facts to memory for an edge in
examinations, and defenders of rote among educators
argue that the effort involved acts as a sort of bench
press for the brain.
The cerebral equivalent of Arnold Schwarznegger’s
bulges, and the discipline it takes to acquire them,
have served Indian programmers well in adapting to the
structured and tightly controlled processes essential
to producing the exceptionally accurate software that
has earned Indian companies CMM Level-5 certification.
Most Western programmers scorn those methods as mental
straitjackets. Frederick Brooks, a revered American
authority on their craft, has captured what they love
best about it, which is software’s rarefied
dimensions: “The programmer, like the poet, works only
slightly removed from pure thought-stuff. He builds
his castles in the air, from air, creating by exertion
of the imagination. The media of creation are so
flexible . . . so readily capable of realising grand
conceptual structures.”
Yet “pure thought stuff” is also a perfect
encapsulation of ancient India’s contributions to the
world’s scientific heritage – which are marked by a
bias towards abstraction unencumbered by empiricism.
Some schools of physics in India developed, in about
600 B.C., atomic theories entirely separate from Greek
atomism, constructions that many Western physicists
scoff at for not being based on experiment but purely
on intuition and logic. But others marvel at how much
closer the imaginative speculations of Brahmin atomic
theory have come to current ideas in theoretical
physics than those of any other pre-modern
civilisation – something no one has ever been able to
explain.
“The Indians advanced astronomy by mathematics rather
than by deductions elicited from nature,” the science
writer Dick Teresi has noted in Lost Discoveries. But
Indian mathematics was also distinctively airy-fairy.
Whereas Greek mathematics was largely extrapolated
from mensuration and geometry, it was abstract number
theory in which the ancient Indians distinguished
themselves. Zero, infinity, negative and irrational
numbers – all concepts that the Greeks dismissed for
centuries as ludicrous – were Indian conceptions.
Spatial extension and quantities of objects were far
less interesting to pioneering Indian mathematical
minds. In fact, the Indian leaning towards abstraction
– so deep-seated that, to this day, theoretical
physicists and mathematicians outrank every other sort
of egghead in status on the subcontinent – explains
India’s relatively poor showing, historically, in more
practical kinds of science. The sinologist Joseph
Needham observed that their study would have entailed
defying Indian caste rules about contamination, and
about contact between Brahmins and artisinal workers.
The progress of ancient Indian knowledge of
physiology, biology and anatomy was held back by the
religious taboo on contact with dead bodies.
It was the supreme pragmatists, the Chinese – whose
intellectual traditions all favoured practicality and
action over airy speculation – who were the
technological geniuses of antiquity. They invented
paper, seismographs, the magnetic compass, the
wheelbarrow, irrigation, ink and porcelain. But
reasoning for its own sake was of so little interest
to them that, unlike the Greeks and Indians, they
never developed any system of formal logic.
It hardly seems accidental that it is through the
mass-manufacture of physical objects – including
nearly every sort of computer hardware, from keyboards
and circuit boards to fully assembled personal
computers – that China is making its mark today and
India, on the ethereal plane of software.
The biggest check on excessive optimism about India’s
economic prospects is the abysmal state of the
country’s material infrastructure – its pot-holed
roads, chronic power failures and foul waterways and
air. Yet, as The Economist reported in February, the
subcontinent’s successes in outsourcing and software
are stimulating critical reforms and progress in
manufacturing industry – a stubborn under-performer
for most of the six decades since Indian independence.
Their power as catalysts is out of all proportion to
the attention-getters’ smallness, in statistical
terms. Information technology accounts for just 3 per
cent of Indian GDP, and in 2002-3, the $9.5 billion
total for software and services exports, including
outsourcing revenues, amounted to less than a third of
Microsoft’s sales of $32 billion in the year to last
June.
Yet, not least because virtually every recent
commentary by domestic or foreign analysts on India’s
future in software has mentioned the infrastructure
problems as a serious bottleneck, improvements have
begun – slowly, but in earnest. If India ever has
smooth roads and lights that can be counted to stay
on, software and outsourcing will deserve a
disproportionate share of the credit.
An incisive observer once remarked about Calcutta that
if only the city’s intractable problems could be
reassigned from the realm of the concrete to the
abstract, keen Indian minds would solve them
overnight. Stretch that idea across the subcontinent,
consider software, and there is a sense in which
cerebration really does appear to be redeeming
obdurate matter.
Cheryll Barron
postgutenberg@gmail.com
Pingback: Why are Indians so good at code? | Matters Mathematical