On the face of it, even the existence of such a movement seems bizarre: Why should large numbers of programmers around the world volunteer to spend hour after hour writing code or catching bugs without the hope of monetary gain? And yet they do. Open-source development has led to the creation of applications such as the Linux operating system and the Apache Web server. Launched eight years ago at the initiative of Linus Torvalds, a Finnish student, Linux-based operating systems have captured nearly 30 percent of the server market.
This success has spawned other initiatives that seek to unleash the creative energy of open-source development. 3Com's PalmPilot is a case in point. Though the PalmPilot is based on proprietary software, 3Com freely encourages the development of compatible games and utilities that can be downloaded and installed in the handset. Microsoft, Apple Computer and Compaq Computer have developed tools for their operating systems through the open-source platform. In 1998 Netscape even released the operating system code for its Communicator 5.0 for public development.
What implications does the open-source movement have for the global software industry? Wharton's Bruce Kogut and Anca Metiu explore that question in a paper titled "Distributed Knowledge and the Global Organization of Software Development." Based on interviews with dozens of software engineers and managers in four countries--the United States, Ireland, India and Singapore--the study points out that the growth of a global infrastructure has made it possible to "exploit globally the opportunities opened by the digitalization of production and products."
Metiu and Kogut see the open-source movement as a tremendous driver of innovation. "The open development model opens up the ability to contribute to innovation," they say. "It recognizes that the distribution of natural intelligence does not correspond to the monopolization of innovation by the richest firms or richest countries. It is this gap between the distribution of ability and the distribution of opportunity that the Web will force companies to recognize and to realign their development strategies." In other words, engineers in China, Israel or India who are unable or unwilling to move to Silicon Valley or the Research Triangle need not be locked out of innovative product development: They can play a vital role in the creation of new products and services.
Metiu and Kogut contrast the open development model with another model of software development commonly used in the industry, which they call the "global project model." Using this approach, software companies set up offices around the world--paying close attention to the locations of their customers--and then develop software wherever it suits them best to do so. This lets the company take advantage of time zone and labor-cost differences to slash software development costs. Example: Trintech, an Irish company that makes electronic payments software, has offices in the United States, Ireland, the United Kingdom and Germany. The company organizes software development work in a way that lets it expand the working day to 16 hours. Some Indian software companies, including Tata Consultancy Services, Infosys and Wipro, boast that since programmers in the United States can hand off tasks each evening to their colleagues halfway in India who are just beginning their working day, they can achieve a "24-hour working day."
The study points out that global sourcing today depends on three main factors: The coordination of modularized tasks, communication, and shared context. All of these have limitations, though. First, coordination in software across borders is particularly difficult when the task is creative. Second, when software is developed internationally, problems can crop up in transmitting certain kinds of knowledge. Third, a shared context limits coordination across distances because people interpret the world in which they live in different contexts.
Metiu and Kogut found that software companies typically outsource only routine tasks to offshore sites and aim primarily at exploiting the low-cost advantage. For example, they looked at a large financial services division of an American company which did only simple subcontracting work with Indian software houses. They also studied another telecommunications company which developed software offshore. When this company's customers demanded features embedded in software rather than in hardware, it was hesitant to commit to outsourcing for strategic products.
However, their research also showed some positive developments that encourage moves by companies toward greater innovation. For example, they found that many offshore software companies like Infosys and Wipro want to be more than just low-cost vendors. They wanted to take part in the innovation process by developing their own products. The researchers believe that global outsourcing will cause global capabilities and aspirations to converge.
A key question to the researchers was: Can companies develop a strategy in which the distinction between onshore and offshore software development merges? At Trintech, for example, the company has tried to make virtual development of software possible by moving closer to its customers in Silicon Valley. The company still has been unable to fully exploit the 24-hour global sourcing cycle. While this model represents a radical departure from the early days of software development, it still did not significantly enhance the scale of innovation.
Kogut and Meitu note that customers also are realizing that even as costs fall, innovations need not occur only in the richest markets. But how is this possible when innovations are designed by the customer, closest to the market? The researchers found that the market itself may move to where the software is being developed. Case in point: Wipro in Bangalore had a customer located in San Diego that requested the development of a software product where the specifications were laid out, the architecture drawn up, and the final product shipped back to the customer. Wipro's software team never met the customer.
So how can companies profit by this strategy? It is possible to do so by shifting revenues from selling the software to the provision of services, online support and training. Further, if companies are able to combine the open-source model with some proprietary software, as Netscape did with its Mozilla licensing policy, they can still maintain a competitive edge.
In all this, an important aspect that Metiu and Kogut studied was why software engineers across the world would volunteer their time to develop free software. Their conclusion: Engineers intrinsically have a gift-giving culture, and the motivations are also driven by norms of reciprocity, reputation enhancement, and by love of their work. They may also benefit financially in the future. Vendors and engineers are increasingly asking to move from a mere fee payment structure to a share of profits in the final product.
Metiu and Kogut conclude that open development is a model that can be transplanted from software to any field where tasks can be broken up into modules and where a wide understanding of the common language and culture exists. They add that "the promise of harnessing the intelligence in the global community through Web-supported innovations will be a key element in the strategies of firms, as well as of public and creative institutions."
To read more articles like this one, visit Knowledge@Wharton.
All materials copyright © 2001 of the Wharton School of the University of Pennsylvania.