If one programmer writes 20 lines of program in a day, how many programmers will it take to write a million lines in one year? The simple-minded answer would be (go on, I’ll work it out for you!) 200 programmers working 250 days a year could write a million lines of program.

Questions like this were considered quite seriously in the optimistic 1960s, when everything was going to be possible using computers. In his book, The Mythical Man Month, Fred Brooks demonstrated that these arguments were based on completely wrong premises. His experience at IBM showed that when programmers worked in large teams, they were, in fact, much less productive and the progress of the project was that much harder to manage. His conclusion was that using a ‘mythical man month’ to estimate how long a programming task will take, or how many people it will need, is a recipe for failure. And that is still true today, whether you call it a man month or a person month.

Flash forward 50 years. Indian software exports zoom to $20-30 billion a year. Growing at a yearly rate of even 25 per cent, how many years will it take for the software industry to earn more than the rest of the Indian economy? And how many people will it need to employ by then?

If the first question makes no economic sense, the second is still being taken remarkably seriously. There are persistent claims that India’s software industry will need between 2-5 million programmers in the next 5 or 10 years. And they ignore the fact that the basis of the estimation is remarkably similar to the one that Brooks destroyed many decades ago.

The simplistic argument also leads to the worry that, with increasing prosperity here, India’s place in the software world can be lost to countries with lower wage levels. After all, surely the lessons that India’s software industry learnt in the last 15 years can now be used to create new competitors, and in far less time? Where better to do this than in lower wage countries with a relatively large educated population ? Vietnam or Cambodia, for example?

Lurking behind this is the conditioning bred of decades of national doubt, coupled with the moralistic belief that those who get to the top by making a quick buck will eventually go the way of all hucksters and charlatans. Despite the middle class rushing to soak up the shares of Indian software companies, there is a deep reluctance to accept one simple fact: Indian software companies have together created a genuine economic force that has forever changed the way that countries make, use and maintain software systems.

If the ‘mythical man month’ is not sufficient reason why simplistic estimates will not work, there is also the fact that the methods of software production are undergoing fundamental changes. There are automated program generation systems that can raise the productivity of programmers by a factor of 10 ?15. This means that far fewer programmers will be needed to complete a programming task. Automated tools will also bring the advice and knowledge of the expert to every programmer, so these few programmers will do a far better job than their predecessors. This is certainly for the better because today’s methods are remarkably error prone.

We do not need to go far to find automated software generation systems. There are already Indian companies producing them to standards matching those now emerging from the market leaders. So tomorrow is here in India today, and here to stay.

That’s the good news. It’s harder to guess what impact this will have on the way software is developed the world over. Indian software companies prospered because they could do the same job as their US or European counterparts more cheaply (and, let’s face it, do it better). That was a compelling argument and, despite the fervor against job losses, US companies that did not outsource their software development came under fire from the analysts. Cisco’s Scheinman put it succinctly: “We came to India for the costs, we stayed for the quality, and we’re now investing for the innovation.” (quoted in Business Week , 22 August 2005).

What if the equation were to change and so few programmers were needed for a task that organizations even in the West were able to afford the costs? After all, there is an overhead associated with outsourcing and this can outweigh the salary arbitrage between the US and India, especially when the number of programmers drops. What if automated software systems become so widely available that the experience and expertise of Indian companies are available through these systems, not just in India but also in Vietnam and Cambodia, for example?

This is where automation stops and skills take over. No automated tool will ever match the capability of a skilled programmer. Nor will it replace the system engineering capabilities of an experienced team. It is here that Indian companies must focus, not just in raising their staff numbers but in providing them with skills that ensure that they can innovate where it matters most.