The biggest challenge for the open-source community is that there are too few open-source developers, according to Michael Tiemann, vice president of open-source affairs at Red Hat Inc.
Tiemann is the latest among a number of key executives from the Raleigh, N.C.-based Linux vendor to visit India, which has a large base of software developers as well as a number of companies that are investing in computerizing their operations. In a telephone interview from Mumbai last week, Tiemann talked to IDG News Service on issues relating to the open-source movement, Red Hat's strategies and his recent blog debate with Jonathan Schwartz, president of Sun Microsystems Inc., on Sun's commitment to open-source. Below is an edited transcript.
Worldwide, are you seeing a greater turnout of developers using open-source? The leading developer regions I have seen have been the U.S. and Europe. But I think that is going to be changing very rapidly. We are getting a lot of news from South America. Brazil, Venezuela, Peru have all either announced or are in the process of announcing mainstream Linux work for government. When Brazil puts their developers on open-source, that is going to be a huge increase.
Are you seeing a lot of participation in open-source development by Indian developers? There is no question that there are Indian developers participating in open-source, and Red Hat certainly has a number of them. We are localizing Fedora [the free open-source operating system project] for six different Indian languages, and we are getting a lot of participation from the community on that. At the same time, I don't see as many people in leading positions in India wanting to be known as open-source developers. Maybe it is a little bit like the way things were in the U.S. maybe 10 years ago. People work for companies, they feel loyal to the company, and what some people don't understand is that one can be loyal to the company and at the same time develop open-source.
What is the biggest challenge for the open-source community? The biggest challenge right now is that there are not nearly as many open-source developers as there could be. The biggest challenge right now is getting more people excited [about open-source development]. The open-source community is a challenging environment to work in. Opinions are very strongly held, and it is not really who you are, but what you can do [that counts]. Some developers respond positively to the meritocracy of open-source, and some do not at all.
What are the management challenges involved when you scale to, say, 100 million open-source developers spread across a number of countries? There is a study from James Herbsleb at Carnegie Mellon University about both open-source and proprietary projects. In this study, 10 to 15 developers are typically responsible for 80% of the project. What that math tells you is that open-source scales by being able to have more and more projects. I don't ever think that there will be a 100 million people working on one library in Linux. Because of the supermodularity of open-source, the ideal resource allocation for a 100 million developers is to be working on, say, 10 million projects.
Would there be a danger of forks in Linux as the number of developers increases? I just don't think it will happen, because we just haven't seen it. Sun claims that they need to keep Java proprietary so that it won't fork. And yet, as proprietary Java, it did fork. IBM came out and released Eclipse, and lo and behold a hundred companies have joined it and there has been no forking on Eclipse.
You have been in the news recently for your blog spat with Jonathan Schwartz, president of Sun Microsystems. In hindsight, what do you think you gained or lost from it? Given the number of comments which people delivered to my blog, I think actually my response had the intended effect, and that is that Jonathan Schwartz for many months has been saying a lot of things that really did require a response. By making one response, it gave the [open-source] community an opportunity to state their frustrations at some of the things he has been saying.
The Linux Standard Base 2.0 of the Free Standards Group attempts to get some standardization into various distributions of Linux. When will Red Hat start shipping LSB 2.0-compliant products? We are really looking towards LSB 3.0, because the 2.0 compromise is not compatible with prior decisions we have made with respect to C++. We have been very successful to certify against LSB Version 1.0, and we will continue to stay compatible with that. We think 2.0 overreached in what it was trying to do. We have it from the LSB people that there will be a 3.0 out in less than six months, and we are quite confident that that version will be adequate for our needs and everybody else's.
As localized versions of Linux take root in countries like China, Malaysia and Thailand, with active promotion of the local governments, how does Red Hat plan to counter that? I was talking with some Indian companies the other day, and they asked me how we were prepared to deal with competition from local Indian companies. I basically said that if the local Indian companies are developing open-source software, according to open-source principles, then we benefit in any case, because we have got the libraries that that software will need.