India’s software industry is growing at 32-33 per cent a year, four times as fast as the rest of the economy which grows at eight-nine per cent. The software services segment of this industry is estimated by the trade body Nasscom to reach $17.5 billion in revenue this year as compared to an industry total of $36 billion, well on its way to crossing $80 billion in 2008 (approximately eight times as much as in 2001). Today, the industry contributes around 4.5 per cent to the country’s GDP and if it continues to grow at this rate, the revenue from this sector would quite soon exceed that from all other sources. Will this make the chief financial officer of a large Indian software company as important as the finance minister? That is unlikely, on any count, so where is the Indian software industry headed for in 10-20 years from now?

It’s not just the laws of economics that are going to force changes in the software industry. Today, the industry is primarily serving the role of an intermediary: providing the technical knowledge and skills to bridge the gap between what computing technology can provide and the services that industry and government must deliver to its users. A bank or an insurance company needs software applications that will enable them to offer customer services that are competitive in feature and cost. A software company delivers its services to such an organisation to build and maintain its software applications.

Competition between software services companies is so intense that their services have become commoditised. It is no longer enough for a software company to just have the right numbers of people with the requisite skills. For some time now, there has been significant pressure to bring down the delivery time and to provide services where they are needed (and through people speaking the right languages). Soon, even that will not be enough.

There is a big move towards evolving common standards across the software industry. Some of these standards will be devised by international bodies, others by the sheer domination of the software products market by large players, and yet others by wide-scale adoption by a major community of users.

For example, the technique of object-oriented design has been standardised by the industry body, Object Management Group, software from Microsoft dominates the software products market and a counter movement by IBM’s Eclipse Framework and the Open Source movement are popular because they allow users to easily share programs.

The next step up the standardisation ladder would be to make it possible for users to build programs out of parts that fit together in standardised ways, much as a component for a car can be replaced by one of another make but with the same specification. A whole software application can then be quickly built by assembling parts or components. A software architect will design a program by specifying the kinds of components that are needed for its various parts, much as an automobile designer specifies the components that are needed for a car. And a small team will source the software components, assemble them together and test the resulting program in weeks, instead of the months and years it may take today.

We could see the whole software services industry moving in this direction in the coming years. And if that happens, where will it take the Indian software industry?

A great deal of what is now being done by this sector for different industries will change beyond recognition. Companies will get differentiated by their roles, either as creators of software components (tomorrow’s software ancillary manufacturers) or as system integrators (the software production line). There will still be a large sector of the industry catering in today’s ‘old-fashioned’ way to the maintenance of legacy software, but that will shrink as new software components are brought in to replace portions of the older systems.

Underlying this will be the increasing automation of software production. Software will be specified in terms of models and systems will automatically generate programs from these models. This kind of automation is already in use in advanced companies and a major change will occur when these techniques are in widespread use.

The other major effect will be on the nature of the people working in the industry. Large armies of programmers will be replaced by people with domain knowledge — that is, knowledge of an industry — and by specialist system integrators who are able to assemble a system quickly and reliably from components. There will be little room for people with average software skills as programs will increasingly be generated automatically. The industry will need specialist software engineers who can meet the pressures for quick delivery, not novices who learn programming on the job.

None of these changes is speculative science fiction. They may not happen quite as suggested here, or quite as soon, but happen they will. If India’s software industry is nimble, it can lead the changes, rather than be overwhelmed by them.

Dr Mathai Joseph is executive director, Tata Research Development and Design Centre, Pune. He earlier was chairperson of software engineering at the University of Warwick for 12 years and a visiting professor at Carnegie-Mellon University and at Eindhoven University of Technology. This is a monthly column.