Linus Torvalds - Just for Fun

It's inspiring that people are using technologies like Linux to just make a better gas pump. That sort of innovation is not likely to have happened within the confines of a company, because if you were a company taking Linux to market you would go for the obvious, which right now is the server market or the high-end desktop market. But open source in general allows companies to make their own decisions about what they want to do. So it's Linux in embedded devices. It's Tivo running Linux and the Transrneta Web Slate running Linux and Telephony using Linux. This is how billions of dollars in wealth is being created from open source.

It's like letting the universe take care of itself. By not controlling the technology, you are not limiting its uses. You make it available and people make local decisions -- to use it as a launching pad for their own products and services. And while most of those decisions don't make sense in the larger scale of things, they actually work really well. This is not about trying to spread Linux. It's about making Linux available and then letting it spread itself. And this doesn't apply only to Linux. It applies to any project that's open.

Open source makes sense.

People don't quibble with the need for free speech. It is a liberty that people have defended with their lives. Freedom is always something you have to defend with your life. But it's also not an easy choice to make initially. And the same is true of openness.You just have to make the decision to be open. It's a difficult stance to take at first, but it actually creates more stability in the end.

Think of politics. If the logic that's used against open source were applied to government, then we would always have one-party rule. Obviously, a single-party rule is a great deal simpler than our system of multiple parties, the open political system in which much of the world operates. With one party you don't have to worry about getting agreement with other people. The reasoning would follow that government is too important to waste on the give-and-take of openness. For some reason people see the fallacy of this argument as it applies to politics and government, but not as it applies to business. Ironically, in business it makes people nervous.

The arguments a company uses to keep from opening up technology are convincing. That's not how things are done , management says. And it's scary. People are frightened of change, partly because they don't know how it is going to turn out. By sticking to the status quo, a company can make a better judgment of where it will go, and sometimes that seems more important than being hugely successful. These are companies that will be predictably successful instead of being unpredictably really, really, really successful.

It isn't easy for a corporation to turn an existing product into an open source product. There are a host of thorny issues. For one thing, over the course of months or years that it developed the product, the company built up a great deal of internal knowledge. This in-house intellectual property is the company's bread and butter. The organization is unwilling to relinquish the intellectual property that keeps it alive. But also, the very existence of this internal knowledge creates a barrier for outsiders. It discourages them from participating.

But I have seen companies make the move from closed to open. One story is Wapit, a Finnish company creating service and support infrastructure for various interactive devices. This project involved the company's wall-phone-style Web server. For them, the decision to open source their software makes perfect sense. They want to build up their service, but they have to build up their infrastructure first. That requires writing a lot of software. It's a necessary evil. So instead of viewing it as a decision to make its intellectual property available to others, they look at it this way: The writing of software takes a lot of engineering time, but it isn't something that creates value from being tightly held at the company.

There were a few things working in Wapit's favor. First, it was not a huge project. Second, the decision to open source was made in the early stages of the company's existence. Management reasoned that it had the resources in-house to develop the product, but it wanted to push to have something more than could be created in-house. It also determined that open sourcing is a good way of furthering Wap as a standard for others to work on.

Early in the game, the company asked for my advice and I told them they needed to fight the urge to have decisions made internally. I suggested that if they were having meetings in which decisions should be made, those meetings should not be closed to outsiders. By keeping the decision-making process a company affair, they would run the risk of alienating outsiders, who would have trouble getting past the company's Old Boy's network. That's one of the major problems of establishing and maintaining an open source project from a corporate environment. It's easy to give lip service to open source, but it can unintentionally degenerate into a two-tier society: Us vs. Them. A lot of decisions get made the easy way -- sitting at the cafeteria table discussing the options and developing a consensus without ever opening up the discussion to the outside. People from the outside who have valid opinions are essentially voted down by the fact that the decision was already made in the company cafeteria.

This was one of the problems that plagued Netscape in the months following that company's much-heralded decision, in the spring of 1998, to release the source code (called Mozilla) for its next-generation browser. It took a long time before the project truly lived up to its open source promise. There was a camp of Netscape insiders who would not accept small patches from outsiders. Everybody inside knew each other, and if they weren't physically sitting around a coffee shop making decisions, there was a virtual coffee shop in which the insiders stayed fairly close. Instead of being seen as embarking on the first great experience in taking an existing commercial project and opening it up, Netscape generated a bit of bad press. When word of its inactivity got out, Netscape could no longer take the high moral ground. That helped them open up to outsiders. Now the project seems to be much more dynamic.

When folks first hear about the possibility of opening up an existing commercial project, they tend to ask the same questions. One question has to do with how people inside the company would feel about the possibility of having an outsider produce work that is better than their own -- and having that so publicly noticeable. I think they should feel great about it, and great that they are getting paid for not even doing most of the work. In that regard, open source -- or open anything , for that matter -- is unforgiving. It shows who can get the job done, who is better. You can't hide behind managers.

Open source is the best way of leveraging outside talent. But you still need to have somebody inside the company who keeps track of the company's needs. That person may not even be the project's leader. In fact, it could be a benefit to the company if someone on the outside takes it over and is doing it for free. It's fine if someone outside is doing a better job. The trouble is, the outsider might lead the project in a direction that doesn't satisfy the company's requirements. So the company must take care of its own needs. The opening of the project might enable the organization to shrink its local resources, but that doesn't mean it can get rid of them. The project could expand to become far bigger than it would have been at a single company. Outside resources make for a cheaper, more complete, and more balanced system, but there's this flip side: The expanded system no longer takes only the company's needs into account. It actually might consider the needs of customers.