The Open Road

Open source is used to playing underdog to incumbent proprietary vendors. What will happen when open source dominates, rather than commoditizes, markets?

I ask because several open-source projects are not far from owning dominant market share in their respective markets. Mozilla's Asa Dotzler reports that Firefox is "on track to easily reach 25 percent of global usage by the end of the year." That may not sound like much, but given that Microsoft has been losing five percentage points of browser market share each year while Firefox gains five percentage points, and it's not hard to imagine Firefox surpassing IE's market share by early 2013.

Firefox isn't alone. Indeed, the Apache Web Server already dominates the Web server market, even despite new entrants to the market, as Glyn Moody highlights.

Linux, for its part, is still only 13.8 percent of the paid server market, while Windows Server still claims 38.1 percent market share, according to IDC. It has a long way to go, but in some markets like cloud computing and the growing Web 2.0 market, it plays a more authoritative role.

So, what happens when these and other open-source projects dominate their respective markets? Will it change how we market open source? Will it mean more research and development dollars must be invested?

Traditionally, open source has done a fantastic job of commoditizing expensive, well-understood markets. While I believe open source can innovate, particularly with companies behind open-source projects, it's still an open question as to whether the financial returns from open-source sales can pay for the heavy R&D and marketing costs that are generally required to create new products and new markets.

Open source has been better at business-model innovation than product innovation, though there are some notable exceptions.

Forget innovation for a minute, however: what will we do when Microsoft, Oracle, etc. are the runners-up, not the market leaders? Microsoft is a convenient (if inaccurate) proxy for all things that are bad in the software world for open sourcerors, but imagine the shift in thinking required to compete when, for example, Firefox has 80 percent market share and IE owns less than 20 percent. Who will we blame for our problems when our straw men are gone?

Perhaps none of this matters, however, as we could see dominant community-led open-source projects fork themselves long before they reach critical, market-dominating mass. It's not hard to imagine splinter groups forming within big open-source projects to take them in different directions, even as Joomla did with Mambo, Ubuntu did with Debian, etc.

The antidote to this is the open-source foundation. Among the examples of strong open-source projects that haven't forked--Eclipse, Apache Web Server, Mozilla Firefox--foundations have been critical to keeping these together. Linux, for its part, has been forked many times, but its core is held together by the Linux Foundation.

I believe the key to attaining dominant market share, and to preventing forks, is the open-source foundation. Over time, I suspect we'll see more "open-source companies" separate themselves into foundations, to manage the code, and corporations, to manage the monetization. This may be the only way to both liberate and dominate at the same time.

As the global economy continues its slide, U.S. politicians have staked out their positions at the stimulus trough, stopping only long enough to blame the other party for the world's problems. Now, more than ever, we need to focus on sound policy, not savvy politics. President Obama ran on the promise of bipartisanship but has come up against the stiff reality of a deep Republican-Democrat divide.

Given such serious divisions, how can we focus our government on policies, not politics? Mozilla, the creator of the Firefox browser, offers some clues.

Yes, that's right. Mozilla. For those who wonder how software could have anything to do with politics, well, you haven't been around open-source software long enough. There's a good reason the tagline for this blog is, "The business and politics of open source." For something as simple as a collection of ones and zeroes, open-source software drags its diverse adherents through a myriad of ideological minefields on a daily basis.

Just think of the bile spewed over the Windows-Linux war, or the antagonism between free-software revolutionaries and open-source advocates, and you get a taste for how deeply political open source can be.

And yet the Mozilla Foundation largely evades divisive jingoism. It just helps create an exceptional browser without mucking through the mire of ideological debates in the process.

The same holds true for the Eclipse and Linux foundations, in their spheres. Each of these foundations provides an opportunity for corporate and community interests to tackle difficult development projects with minimal rancor and little politicking, at least as exposed to the general public's view.

I've noted before that I think open-source foundations provide a clue as to one bright future for open source, but I'm also starting to wonder if they could provide a clue as to how to enable President Obama in the U.S., and other government executives across the globe, to encourage collaboration toward resolution of difficult, seemingly intractable policy problems.

Much of the problem with politics is that the information that feeds into political debates is suspect because it is so biased. As in computer science, garbage in, garbage out. Information is the "source code" of political debate, yet its nutritional content is sorely lacking here in the States.

Here's how Mozilla and open-source foundations provide a clue as to how to improve things. I liken Mozilla to a Rand of sorts: an organization dedicated to discussing and resolving difficult problems with little concern for who is right.

It may sound utopian, but I wonder if, in similar manner, President Obama couldn't require some form of Rand-esque analysis in congressional disputes to serve as a primary information source, rather than leaving it to competing special interests to funnel biased information into the Congresspersons of their choice?

Or, more broadly still, maybe what the federal government needs is more open caucuses, in which the big, thorny issues are discussed and debated. Perhaps it would help if the people stumping for votes aren't the ones in the debate: perhaps they would simply be consumers of the information, rather creators of it.

This, after all, is how the Linux Foundation and the Eclipse Foundation function. Big, partisan corporate interests fund these foundations, and then largely get out of the way, while developers, sometimes subsidized by this corporate funding, write the code. Congress could do the same.

What we do matters much more than who gets credit. If Congress will fund "open-source foundations" to create information which, in turn, will fuel its votes, perhaps we'd be one step closer to intelligent policies, rather than inane politics.

Too idealistic?

Follow me on Twitter at mjasay.

I came across news this afternoon about the LiMo Foundation endorsing the Open Mobile Terminal Platform specification, and I realized that I didn't care. It's probably big news, but I couldn't get excited.

I feel the same way about most things that come out of the Apache Software Foundation (ASF), but probably because the ASF isn't one for making big announcements (except when it collects cash from Microsoft). Even so, I know the ASF is a hugely important organization. Its only "problem" is that it lacks an active public relations team. But I doubt it sees this as a problem.

The Symbian Foundation, too, shows promise, but it has also been somewhat quiet.

Now contrast these organizations with the Linux Foundation, Eclipse, and Mozilla. All three provide effective, though different, examples of what an open-source foundation should be doing. They're active. They're noisy. They're influential and even critical to the development of key open-source projects.

What do they have in common? Well, each has strong leaders, both technical and business management. Each has a limited but important mission. (Here's the Mozilla Foundation's mission.) Each has the trust and funding of key constituents that contribute both code and cash. And each publicly advocates their projects.

I've suggested before that Eclipse offers a vision of open-source development to come. If anything, this belief grows daily, but it's not just about Eclipse. Foundations offer a way to make open-source development more corporate (organized in such a way that commercial vendors can participate with fewer reservations) without becoming commercial, a turn-off for many would-be code contributors.

Understanding how successful open-source foundations function, and why they succeed while lesser peers fail, is therefore critical to understanding the future of open-source development. Mozilla, Eclipse, and the Linux Foundation offer glimpses of good things to come. As for the other open-source foundations...I think they still have work to do.

OStatic's Sam Dean speculates that 2009 could be the year of open source mergers and acquisitions, what with the super-low valuations of Sun, Novell, and Red Hat. He may be right, though we might also see these (and others) snapping up the low-hanging fruit of even smaller open-source companies in an attempt to unify a growing open-source ecosystem under one corporate roof.

As I fell asleep last night, however, a different thought struck me: do Eclipse and Mozilla suggest an entirely new form of M&A for open source?

Yesterday I suggested that creating an OpenOffice foundation might be the best way to resolve its alleged management problems. Indeed, as I've argued before, foundations around open-source projects may well be one of the best ways to grow open-source projects' relevance and influence.

Could this be a more efficient way to provide industry benefit without trampling on developers' benefits? Consider the alternatives.

The increasingly normal (commercial) route goes as follows: Project X is started by a developer in her spare time. Project X grows due to great code and solid leadership. VCs notice this growth and invest in a company (Company X) around the project (e.g., Acquia for Drupal, Digium for Asterisk, etc.). Company X competes with IBM et al. but IBM et al. wish that Company X would go away. In some cases, IBM et al. acquire Company X, and capture its value for themselves, leaving the rest of the industry and, often, the project, a bit moribund.

What would happen, however, if the industry had a mechanism for allowing interested corporate parties to provide an exit for Project X's (or Company X's) core team? Instead of selling to one company, in other words, Project or Company X would sell to the industry, as it were, and would become Foundation X, with its value would becoming industry property.

There would be huge benefits to the industry with such an approach, and solid returns for the developers at the open-source project. The problem, of course, is coordinating the resources necessary to purchase an exit. Given corporate gamesmanship, this is likely to prove intractable.

It's too bad. My own company has had large corporate interests approach Alfresco to change how we operate to make us more of an industry standard, but at our expense, not theirs. I imagine, too, that projects like JBoss and MySQL would have made amazing foundations, but I doubt their founders (and their investors) would have been willing to forgo their $350 million and $1 billion paychecks, respectively.

It's not enough to demand that Sun turn OpenOffice into a foundation, or to howl at the gates for IBM to open source its Notes and Domino products. We can't rely on big companies to spend money to develop assets, and then release them out of charity.

Even so, there must be some way to make foundations into a profitable exit for development teams. As an industry, we'd be the richer for having more foundations, if only we could figure out how to help developers become richer in the process and to coordinate the corporate resources necessary to make it feasible.

The Eclipse Foundation has released an updated roadmap, one that recognizes Windows current importance, but also sees Windows Vista as an opportunity to nudge developers to the Eclipse platform, potentially away from Microsoft's Vista:

Approximately 85% of Eclipse download requests are for the Windows OS. With the Vista release there are a number of efforts to port Windows applications. This presents an opportunity for organizations who will take the opportunity to migrate to the more ubiquitous and portable Eclipse platform. In order to leverage the opportunity as much as Vista, it is essential that relevant Eclipse projects support and leverage VISTA. For example, Avalon APIs need to be implemented in SWT [Eclipse Standard Widget Toolkit].

As The Register's Gavin Clarke suggests, "SWT potentially lets applications run on Linux, Mac OS X, and different versions of Windows," and so broadens Eclipse's appeal (especially for Rich Internet Applications), while potentially diminishing Vista's.

Microsoft, of course, has started working with Eclipse to make Eclipse a "first-class platform on Vista." It's too soon to tell if Eclipse will succeed in its apparent attempt to downplay Vista as a development platform, but it's interesting to see this interplay between coopetition and competition going on between allegedly neutral Eclipse and allegedly not-so-neutral Microsoft.

CNET's Stephen Shankland has already picked up on Google's decision to allow two popular open-source licenses back onto its Google Code open-source repository. Up until now, the Mozilla Public License (MPL) and Eclipse Public License (EPL) were both banned from the site.

The reasoning Shankland reports for Google belatedly approving the licenses, however, is a bit bizarre. In the case of the EPL, Google's Chris DiBona argues:

Eclipse is an important, lively, and healthy project with an enormous plug-in and developer community that uses an otherwise duplicative license. They aren't interested in using the BSD or other open-source licenses that are readily combinable with EPL code. We have decided that after nearly 2 years of operation, that it was time to add the EPL and serve these open-source developers.

Well, yes. But that was true before Google opted to ban them. DiBona is a smart, super open source-savvy guy. He didn't need anyone to tell him that the EPL is critical to Eclipse, and that Eclipse is highly important to open source. So why the lag?

As for the MPL, while DiBona doesn't state it outright, I suspect that Google's decision to re-up its commitment to Mozilla for three more years probably involved some strained discussions about Google's weird decision to dump the MPL, one of the industry's most popular open-source licenses.

Regardless, all is well that ends well. Google came to the right decision, however odd the logic.

I received an interesting email today from the Eclipse Foundation, interesting because it cut against an experience I had at work, and against Red Hat CEO Jim Whitehurst's evangelism to convince more enterprises to contribute to open-source software projects.

I'm aware of a number of mission-critical applications happily running on open-source software, unpaid and unsupported. Fortunately (or unfortunately, depending on how I feel about my quarterly numbers at any given moment :-), my own company's code is included in this observation. In each case, the enterprise contributes nothing back to the respective open-source communities.

Eclipse, however, is having a different experience:

Consumers of technology -- not just vendors -- are now getting involved in Eclipse. For example, Deutsche Post contributed code to start the Eclipse Swordfish project (SOA runtime), Boeing contributed its homegrown software to start the Open Systems Engineering project, and VA Hospital and Kaiser Permanente are driving the Open Healthcare Framework.

This is fantastic. Such contributions still represent a tiny sliver of overall Eclipse contributions, but it's a start. IBM kicked off Eclipse six years ago with a large code drop, and remains one of its biggest backers. The Eclipse Foundation now has more than 185 members. Hopefully, in the not too distant future we'll see companies like Citigroup joining the Eclipse Foundation and contributing code there.

Eclipse works because it doesn't need much in the way of cash to operate. Some cash is critical to fund its operations, but code is the more important currency for Eclipse. As I've written before, does Eclipse point to the future of open-source development? Is there a lesson in this for commercial open-source vendors, as Kris Buytaert suggests?

CIO.com suggests that open source may be the future of enterprise innovation, echoing the Bank of New York Mellon's comments on the subject last week. The question is not why use open source, but how to best use open source.

Riffing off the Eclipse Foundation's Mike Milinkovich, CIO.com writes:

[I]f you develop in an open source model and other companies adopt what you develop, you have a higher chance of longevity in the code base. In other words, you can develop a custom solution to a unique business problem with less fear that your solution will turn into a dead-end legacy system when things, inevitably, change in a few years. Milinkovich also sees more CIOs banding together with industry peers to develop common open-source solutions to standard industry processes, thereby saving money by sharing costs and ensuring interoperability.

Amen. Proprietary software is all about vendor-driven innovation. Open source is all about user-driven innovation, collaboratively married to vendor innovation. It's a two-way street, and it's much, much better for both vendors and users.

eWeek has put together a solid list of the top-15 open-source business influencers in the industry today. It's much the same that I would have devised had I come up with such a list. Names like Linus Torvalds (Linux), Mitchell Baker (Mozilla), Mike Milinkovich (Eclipse), and Larry Augustin (most startups known to humankind) make the list. Also John Roberts of SugarCRM, who was really the one who made commercial open source as big a topic as it is today, at least beyond Linux and middleware.

But the list is also notable for its inclusion of some people that might normally escape public notice, but who deserve it all the same. Tim Golden (Bank of America) is one such person. Another is Peter Fenton (Benchmark Capital). Few recognize just how deeply involved Peter is in the open-source venture market.

Others on the list, like Mark Shuttleworth, are well-known but not yet for the right reasons. Mark is known more for his engineering side (Ubuntu), but I think his most lasting impact will be with the uncompromising business principles that he brings to Canonical.

At any rate, good people getting the credit they're due.