Learning from Mozilla’s mistakes

– By Robin “Roblimo” Miller –

When the Mozilla project started, it immediately became the number one poster child for Open Source software development. Now its luster is tarnished to the point where closed source advocates point to Mozilla as an example of how Open Source cannot compete one-on-one with proprietary software, in this case with Microsoft’s Internet Explorer. Is this true? Or was Mozilla’s development process, not the fact that it was Open Source, to blame for its problems?Netscape was one of the first examples of dot-com excess, with liberal work rules that allowed employees to bring dogs to work, Nerf gun battles at all hours, and a generally quirky corporate style. From all reports, working at Netscape during the glory years was so much fun that some employees spent all of their waking hours there.

Sadly, while Netscape employees were engaging in company-sponsored play that liberated their creative juices, Microsoft’s developers were grinding on, perhaps without passion, making Explorer a little better and more usable, one step at a time, day after day, month after month.

Then Netscape decided to throw open their development process so thousands of eager volunteer programmers all over the world could participate, hoping this would lead to a surge in productivity by creating a mass coding frenzy that would make more software faster than it had ever been made before.

At the time, I was suspicious of Netscape’s motives: Were they going in the Open Source direction because they truly believed in sharing, or were they trying to … exploit is a harsh word … the efforts of Open Source believers to further their own corporate fortunes?

The idea of having people work for free is certainly wonderful from the viewpoint of a corporate executive whose primary duty is to enhance shareholder value. Anyone who thinks this kind of thinking isn’t at least partially behind many companies’ adoption of Open Source has a shaky grasp on reality. Do you think IBM would embrace Linux to the tune of one billion dollars if the bean counters there didn’t figure that investment would either save two or three billion dollars worth of internal development costs or create many billions of dollars in sales IBM wouldn’t otherwise get?

So there was Netscape, suddenly a division of AOL, asking software developers to donate their time to what was, at heart, a corporate project. I asked Netscape PR people more than once, back then, how many outside developers, as opposed to Netscape employees, were working on Mozilla. Not once did I get a concrete answer. I asked for interviews with internal Mozilla project leaders, but got no replies to my emails and phone messages. Finally I gave up. Perhaps the project itself was open, but the corporate structure behind it certainly wasn’t.

I have tried a number of Mozilla builds along the way, but so far I am sticking with tried-and-true Netscape Communicator 4.7X, despite its flaws, as my primary Web browser and email client. I have tried Konqueror and Opera and all the rest, but old-fashioned Netscape still fills my needs best. Yes, I would like to have Mozilla and all its promised stability and new features. But a fully work-usable Mozilla seems to keep sliding farther into the future.

Meanwhile, Microsoft’s Internet Explorer has gotten pretty darned good. Note that AOL, Netscape’s owner and Mozilla’s sponsor, is only now even considering the Mozilla-generated Gecko rendering engine for use in a default browser, and even then only for its much-smaller-than-AOL CompuServe division.

(The latest AOL software, just released, uses Explorer as its default browser.)

I have watched other Open Source projects progress at amazing speed without direct corporate sponsorship. I have watched eager volunteer developers contribute endless hours of time, and I’ve produced a number of detailed bug reports and helped with documentation on several projects. I have no problem contributing to the community. If anything, I believe it is my duty to do this because I use plenty of Open Source and Free Software, and I regard the time I spend helping Open Source development, wherever I can, as my personal in-lieu-of-cash payment for the “free as in beer” software I use every day. But I am not going to work for AOL for free.

There is nothing inherently wrong with corporate sponsorship of an Open Source project, but if that project is going to attract a large number of volunteers, it must not be seen as little more than a chance to help the sponsoring corporation make money. Sun seems to have figured this out with their OpenOffice project, which they say “… will be administered by the OpenOffice.org Foundation,” a promise weakened greatly by the disclaimer, “At this point we are working on the concrete structure and agenda for the foundation.”

An endowed foundation, not directly controlled by the company that funds it, may be the best and purest way to funnel corporate support to an Open Source project. It removes most of the odor of corporate greed, and if the foundation uses one of the widely accepted “generic” Open Source or Free Software licenses, it removes suspicions that taint corporate lawyer-generated licenses like the Mozilla & Netscape Public Licenses or the only slightly less complex dual licensing scheme Sun uses for OpenOffice.

I respect IBM, Netscape/AOL, and Sun for their attempts to move toward Open Source. It is not easy to give up corporate control over intellectual property, especially for companies whose fortunes are based on patents, copyrights, and other methods of preventing competitors and potential competitors from using their work. But “sort of” Open Source or Free Software development doesn’t seem to work as efficiently as the pure, unadulterated, variety — or necessarily as well as straight-up proprietary software development, either.

What if, instead of trying to keep Mozilla development under their direct control, Netscape had set up and funded an independent non-profit foundation to develop Mozilla from the start? I suspect that if this had happened, we would all be using a reliable, full-featured, tightly-coded 2.X version of Mozilla today, and if that foundation’s output had carried a BSD-style license, AOL would have been able to add proprietary code to the Open Source base to create a default browser for their customers that would be much better than Internet Explorer.

To use a trite phrase, “Hindsight is always better than foresight.” Netscape was a true pioneer in corporate-sponsored Open Source development and had no real past experience to go on. Mistakes were — inevitably — made.

But another trite saying says, “Those who don’t study history are doomed to repeat it.”

Critical analysis of current and past Open Source development successes and failures is not always fun, especially for the people and companies whose projects are being analyzed. But the lessons we learn by looking at high-profile Open Source projects through clear — even jaundiced — eyes, instead of through rose-colored glasses, are more important to the future of Open Source and Free Software than any amount of boosterism for a single project, even one with as many hopes riding on it as Mozilla.