For 12 years, Robert Lefkowitz has been an IT director for major Wall Street firms. During the spring of 2002 he was named director for open source strategy at Merrill Lynch. This spring, the group was scuttled. But when Lefkowitz showed up at the Open Source Convention in Portland in early July to give a couple of talks about the changing role within giant corporate IT departments of open source software-programs in which the code is shared among developers rather than kept a proprietary secret-his star was clearly on the rise: he came away from the meeting with a fistful of new job offers. Leaving Portland, he was set to hand in his resignation on Wall Street then fly to Ottawa, Canada, to join a select list of kernel hackers at a Linux summit meeting. He was going there to tell the Linux elite what features should go into the operating system that would most benefit large corporate users. The MIT electrical engineering grad is one of the few people with both the technical chops and a detailed grasp of enterprise computing to make that kind of presentation-which is why he is fast becoming a prominent figure in the open source community.

Good move. Lefkowitz came away from the meeting with a fistful of new job offers. Leaving Portland, he was set to hand in his resignation on Wall Street then fly to Ottawa, Canada, to join a select list of hackers at a Linux summit meeting. He was going there to tell the Linux elite what features should go into the operating system that would most benefit large corporate users. The MIT electrical engineering grad is one of the few people with both the technical chops and a detailed grasp of enterprise computing to make that kind of presentation-which is why he is fast becoming a prominent figure in the open source community.

Lefkowitz has been mightily impressed by the idealism of open source developers. "They have an idea how they want to change the world to make it a better place. And they're working towards that goal. I love to be around people like that." Lefkowitz spoke with freelance writer Brad Stenger.

TR: What led you to look at the relationship between money and open source software?
LEFKOWITZ: There's the question of how does money flow into the open source ecosystem, and what kind of an impact does open source have on the commercial world. People start out thinking that they'll save on licensing costs because the software is free. But it turns out that more of the financial benefit comes from the network effects associated with the transfer of knowledge, and the interactions that happen between people. The first proposal that I made about getting involved with open source was a recruiting argument: You get the chance to see a programmer's work before you even interview him or her. My pitch was: If these programmers' work is any good, think of what you could save on recruiting fees.
 
TR: Is the open source ideology attractive enough that it could be a magnet to draw individuals in from corporate computing?
LEFKOWITZ: Yes, I think so. You can view open source as a form of continuing education. Many of my colleagues are adjunct professors. It's beneficial to the students to have practitioners involved. It's beneficial to the practitioners to have the interactions with the students. It's good for everybody because it raises the collective awareness of technological solutions. How can you object to that?

TR: So you see open source as akin to academia?
LEFKOWITZ: Yes. One of the things that open source projects should aspire to is to create the definitive literature for a particular problem domain. The distinction between what one's doing in open source and what one's doing commercially would be the same distinction that one makes between academia and business. Spinoffs of good ideas that originate in academia turn into companies. There's room to push forward both the academic version, which is focused on clarity and exposition, and the commercial version, which is focused on scalability and performance.

TR: Who are the customers for open source products?
LEFKOWITZ: The customers are people who want to use computers, but don't want to be developers, or who don't want to be terribly technical. Now, the common view of it is that open source is for the alpha geeks-people who want to fiddle around in the low levels. That's one end, but it's not the whole story. Open source has permeated a huge range of products. Apple and Microsoft have used open source technologies. Oracle uses them. IBM uses them. It's for everybody.

If you're looking at it more from a community and education angle, open source is used in academic settings to teach how computers work. That kind of activity is going to be open source forever:  you're not going to be able to teach somebody how a computer works unless you can show them some code. To the extent we can equate the idea of open source with the idea of education about computers and software in general then the question devolves to:  Who is education important for? And the answer to that is: Everyone.

TR: What about services based on open source software?
LEFKOWITZ: I'm a big believer in incremental improvement-what the Japanese call kaizen. In that view of a service ecosystem you continue to reflect on how you can provide the service better and think about what things you could streamline through automation.

TR: How can open source software keep the innovation premium in place-that is make sure the innovators are rewarded for the risk they take?
LEFKOWITZ: That's a really complicated question, and my answer to it keeps evolving-in part because open source means so many things. Oftentimes if I contribute something to the open source community, it's because I believe that the value to be derived by contributing exceeds the value of not doing so.

TR: What is an example of that?
LEFKOWITZ: People typically do that if you're just looking at the code for reasons related to testing and porting and internationalizing. If you want to port the code to 20 different platforms then you need to acquire 20 different platforms to test it. Small developers can't afford to do that. But if they open source it, other people will do that work.

And then, there's the secret sauce for which you wish to capture an innovation premium. You can take that open source stuff, add your secret sauce, and create some very nice commercial products. Apple does it. Microsoft did it with ActiveDirectory. The advantage is this: if somebody else then wants to take that same code base, and add a proprietary innovation that you don't have, they'll be able to monetize that. It depends on the extent to which you are interested in encouraging a marketplace. Some open source projects, like those based on the GNU General Public License (GPL), are more directed at encouraging the growth of freely available software. Others, like BSD and Apache, are concerned with capturing a financial premium from innovation.

TR: Can you see a future in which open source permeates corporate IT culture? Can open source be a guiding principle in corporate decision-making and strategy?
LEFKOWITZ: A lot of executive suites are focused on collaboration and openness and free exchange of ideas within their companies. They want to encourage their people to learn more, to interact more, to share ideas more within the corporation. Well, for open source companies like Collabnet and Sourceforge, that's their business-going into companies and selling them the service, using open source tools, or proprietary tools, to help them organize themselves and behave internally the way the open source community behaves among itself. That kind of culture is desirable. People recognize that today.

Copyright Technology Review 2003.