New Delhi: In an economic examination of the software realm, a new book titled The Comingled Code comes to an even-handed assessment of the future: The way forward is not open-source or proprietary software, but open-source and proprietary software. One of the book’s authors, Josh Lerner, a Harvard Business School professor, talked to Mint about how economic models can change in the software realm. Edited excerpts:
Could you give an example of an economic precept that breaks down when applied to open-source?
There’s something called the Arrow problem, after Kenneth Arrow, the economist—that we need to have intellectual property to get innovation. No one is going to innovate unless they have the incentive of knowing they’re going to get monopoly over it. The classic example is pharmaceuticals. Without intellectual property, drug companies aren’t going to go develop a new drug, because it’s so easy to imitate it.
The counter argument is open-source. You have all this innovation, yet if you think about the incentives people have, there’s very little. A lot of people are not being paid to develop this. So you’re getting innovation without any of the distortions associated with intellectual property. You’re saying: “This is not the way it’s supposed to work. Yet it seems to be very much happening along these lines.”
Is the image of the open-source tech-geek, working out of altruism or pure interest, still valid?
That’s where open-source started. As the business has matured, it’s changed in several ways. You’re seeing more involvement of for-profit companies, taking their programs and opening up the code— (not) their crown jewels, but some interesting program that still has some challenges. Some examples are open-source companies in Silicon Valley, where open-source is the core and they’re adding proprietary stuff around it. When you go to Sendmail’s offices, they’ve got Sendmail.com on one door and Sendmail.org on the other door. One is the open-source project and the other sells all the enhancements for corporate users who want more documentation, higher-ended functionalities, and some sort of warranty that, if things go bad, they’ll fly a team of people there to fix it.
So is the incentive model in open-source different from elsewhere?
Certainly, we don’t see a lot of this kind of pure recognition in a lot of other places. If you come up with an idea, at Ford, for a better windshield wiper, Ford is not going to put name on the wiper, saying: “Thanks for the contribution.” So it’s a little unique in the way recognition can translate into economic opportunities. You see echoes of it elsewhere. If you develop a new surfboard, you often get some fame, (or) get a board named after you.
How does innovation then work in a realm with an altered incentive system?
This is an interesting question. We have this idea that incentives translate into innovation: The more you pay people, the more they’re going to do. But it’s a little more complicated than that. I’ll talk about one example of another study in a different realm. A couple of guys at MIT (Massachusetts Institute of Technology) compared two ways biologists get grants. The US government tends to give grants for two-three years, and they make very specific goals, and you have to go and deliver. Then there’s the Howard Hughes Medical Institute, which gives grants for 10 years, where you can be more open-ended.
If you think about incentives, the first has a tighter incentive scheme: You’ve got to deliver. The other is more open-ended: There’s not much of a link between what you’re getting paid and your performance. Yet if you look at important innovations, a lot more comes out of the Howard Hughes model. So the rules that apply to motivating salesmen don’t quite carry over to innovation.
You compared case studies across the developed and developing world. How do the dynamics of open-source change?
There are several differences. One is a need for customization. In many developing countries, there’s a sense that proprietary software is not really set up for their needs. The second is the importance of cost considerations; for many users in developing countries, they aren’t able to spend a huge amount on proprietary software.
An issue cutting the other way… is the scarcity of programming talent. Good programmers are in very high demand (in developing countries). So you may not end up in the sort of dynamic that you have in the US, where a lot of good programmers sitting around with a lot of free time. We did a case on HackerTeen in Brazil. In Brazil, there was a real shortage of programmers, so they trained teenagers in programming open-source, to create of a supply of people to work on projects, and also open a door for these kids in terms of upward mobility.
You advise that governments shouldn’t necessarily choose open-source by default. Why is that, especially since if their budgets get leaner, and if they have the resources to ensure that their open-source systems function smoothly?
There’s a tendency among users to look at the purchase price and say: “Well, this one’s zero. It wins.” But…there are all these costs associated with implementation, upgrades, various things.
If open-source wins in terms of total cost of ownership, go for it. But if you look at Brazil a few years ago, there was a dichotomy between politicians and the guys actually doing the IT implementation. The IT guys were like: “Some things we want to use open-source for, some we want to use proprietary software for.” The politicians were like: “Open-source is good. Let’s go with that.” Brazil has become more nuanced since then, but…this is a realm where economic decisions have been made on the basis of emotion rather than data.