Free for All - Peter Wayner (the false prince series .txt) 📗

rather spend their time writing code than juggling drivers on their machine. Good distributions like Red Hat, Slackware, Debian, FreeBSD, and SuSE made it possible for everyone to get their machines up and running faster.

There's no reason why the latest push into the mainstream is going to be any different. Sure, Red Hat is charging more and creating better packages, but most of the distribution is still governed by the GPL. Whenever people complain that Red Hat costs too much, Bob Young just points people to the companies that rip off his CDs and charge only $2 or $3 per copy. The GPL keeps many people from straying too far from the ideal.

The source is also still available. Sure, the corporate suits can come in, cut deals, issue press releases, raise venture capital, and do some IPOs, but that doesn't change the fact that the source code is now widely distributed. Wasn't that the goal of Stallman's revolution? Didn't he want to be able to get at the guts of software and fix it? The source is now more omnipresent than ever. The corporations are practically begging folks to download it and send in bug fixes.

Of course, access to the source was only half of Stallman's battle. A cynic might growl that the corporations seem to be begging folks to do their research, testing, and development work for them. They're looking for free beers. Stallman wanted freedom to do whatever he wanted with the source and many of the companies aren't ready to throw away all of their control.

Apple sells its brand, and it was careful not to open up the source code to its classic desktop interface. They kept that locked away. Most of the source code that Apple released is from its next version of the operating system, Mac OS X, which came from the folks at NeXT when Apple acquired that company. Where did that code come from? Large portions came from the various free versions of BSD like NetBSD or Mach. It's easy to be generous when you only wrote a fraction of the code.

Ernest Prabhakar, the project manager for Apple's first open source effort known as Darwin, describes the tack he took to get Apple's management to embrace this small core version of the BSD operating system tuned to the Macintosh hardware platform.

"The first catalysts were the universities. There were a lot of universities like MIT and University of Michigan that had some specialized network infrastructure needs," he said.

"We realized that the pieces they're most interested in are the most commoditized. There wasn't really any proprietary technology added that we had to worry about them copying. There are people who know them better than we do like the BSD community. We started making the case, if we really want to partner with the universities we should just open the source code and release it as a complete BSD-style operating system.

"We wanted people to use this in classes, really embed it in the whole educational process without constraining teaching to fit some corporate model," he finishes.

Of course, Prabhakar suggests that there is some self-interest as well. Apple wants to be a full partner with the BSD community. It wants the code it shares to mingle and cross-pollinate with the code from the BSD trees. In the long run, Apple's Darwin and the BSDs will grow closer together. In an ideal world, both groups will flourish as they avoid duplicating each other's efforts.

Prabhakar says, "This reduces our reintegration costs. The ability to take the standard version of FreeBSD and dump it into our OS was a big win. Prior to doing the open source, we had done a small scale of givebacks."

This view is echoed by other companies. IBM is a great hardware company and an even greater service company that's never had much luck selling software, at least in the same way that Microsoft sells software. Their OS/2 never got far off the ground. They've sold plenty of software to companies by bundling it with handholding and long-term service, but they've never had great success in the shrink-wrapped software business. Open source gives them the opportunity to cut software development costs and concentrate on providing service and hardware. They get free development help from everyone and the customers get more flexibility.

Sun's Community Source License is also not without some self-interest. The company would like to make sure that Java continues to be "Write Once, Run Anywhere," and that means carefully controlling the APIs and the code to make sure no idiosyncrasies or other glitches emerge. People and companies that want to be part of the community must abide by Sun's fairly generous, but not complete, gift to the world.

The company's web page points out the restriction Sun places on its source code fairly clearly. "Modified source code cannot be distributed without the express written permission of Sun" and "Binary programs built using modified Java 2 SDK source code may not be distributed, internally or externally, without meeting the compatibility and royalty requirements described in the License Agreement."

While some see this clause as a pair of manacles, Bill Joy explains that the Community Source License is closer to our definition of a real community. "It's a community in a stronger sense," he told an audience at Stanford. "If you make improvements, you can own them." After you negotiate a license with Sun, you can sell them. Joy also points out that Sun's license does require some of the GNU-like sharing by requiring everyone to report bugs.

Some customers may like a dictator demanding complete obeisance to Sun's definition of Java, but some users are chaffing a bit. The freedom to look at the code isn't enough. They want the freedom to add their own features that are best tuned to their own needs, a process that may start to Balkanize the realm by creating more and more slightly different versions of Java. Sun clearly worries that the benefits of all this tuning aren't worth living through the cacophony of having thousands of slightly different versions. Releasing the source code allows all of the users to see more information about the structure of Sun's Java and helps them work off the same page. This is still a great use of the source code, but it isn't as free as the use imagined by Stallman.

Alan Baratz, the former president of Sun's Java division, says that their Community Source License has been a large success. Sure, some folks would like the ability to take the code and fork off their own versions as they might be able to do with software protected by a BSD- or GNU-style license, but Java developers really want the assurance that it's all compatible. As many said, "Microsoft wanted to fork Java so it could destroy it."

Baratz said, "We now have forty thousand community source licensees. The developers and the systems builders and the users all want the branded Java technology. They want to know that all of the apps are going to be there. That's the number-one reason that developers are writing to the platform." Their more restrictive license may not make Stallman and other free software devotees happy, but at least Java will run everywhere.

Maybe in this case, the quality and strength of the unity Sun brings to the marketplace is more important than the complete freedom to do whatever you want. There are already several Java clones available, like Kaffe. They were created without the help of Sun, so their creators aren't bound by Sun's licenses. But they also go out of their way to avoid splitting with Sun. Tim Wilkinson, the CEO of Transvirtual, the creators of Kaffe, says that he plans to continue to make Kaffe 100 percent Java compatible without paying royalties or abiding by the Community Source License. If his project or other similar ones continue to thrive and grow, then people will know that the freedom of open source can be as important as blind allegiance to Sun.

These corporate efforts are largely welcomed by the open source world, but the welcome does not come with open arms or a great deal of warmth.

Source code with some restrictions is generally better than no source at all, but there is still a great deal of suspicion. Theo de Raadt, the leader of the OpenBSD project, says, "Is that free? We will not look at Apple source code because we'll have contaminated ourselves." De Raadt is probably overreacting, but he may have reason to worry. AT&T's USL tied up the BSD project for more than a year with a lawsuit that it eventually lost. Who knows what Apple could do to the folks at OpenBSD if there were a some debate over whether some code should be constrained by the Apple license? It's just easier for everyone at OpenBSD to avoid looking at the Apple code so they can be sure that the Apple license won't give some lawyers a toehold on OpenBSD's code base.

Richard Stallman says, "Sun wants to be thought of as having joined our club, without paying the dues or complying with the public service requirements. They want the users to settle for the fragments of freedom Sun will let them have."

He continues, "Sun has intentionally rejected the free software community by using a license that is much too restrictive. You are not allowed to redistribute modified versions of Sun's Java software. It is not free software."

16.1 FAT CATS AND ALLEY CATS

............................

The corporations could also sow discord and grief by creating two different classes: the haves and the have-nots. The people who work at the company and draw a salary would get paid for working on the software while others would get a cheery grin and some thanks. Everyone's code would still be free, but some of the contributors might get much more than others. In the past, everyone was just hanging out on the Net and adding their contributions because it was fun.

This split is already growing. Red Hat software employs some of the major Linux contributors like Alan Cox. They get a salary while the rest of the contributors get nothing. Sun, Apple, and IBM employees get salaries, but folks who work on Apache or the open versions of BSD get nothing but the opportunity to hack cool code.

One employee from Microsoft, who spoke on background, predicted complete and utter disaster. "Those folks are going to see the guys from Red Hat driving around in the Porsches and they're just going to quit writing code. Why help someone else get rich?" he said. I pointed out that jealousy wasn't just a problem for free software projects. Didn't many contract employees from Microsoft gather together and sue to receive stock options? Weren't they locked out, too?

Still, he raises an interesting point. Getting people to join together for the sake of a group is easy to do when no one is getting rich. What will happen when more money starts pouring into some folks' pockets? Will people defect? Will they stop contributing?

Naysayers are quick to point to experiments like Netscape's Mozilla project, which distributed the source code to the next generation of its browser. The project received plenty of hype because it was the first big open source project created by a major company. They set up their own website and built serious tools for keeping track of bugs. Still, the project has not generated any great browser that would allow it to be deemed a success. At this writing, about 15 months after the release, they're still circulating better and better beta versions, but none are as complete or feature-rich as the regular version of Netscape, which remains proprietary.[^11]

[11]: At this writing, version M13 of Mozilla looks very impressive. It's getting quite close to the proprietary version of Netscape.

The naysayers like to point out