Opera's open source conundrum

Suppose there was a software category so ubiquitous that virtually everyone used it and anyone could get the software for free. Suppose, also, that the software was highly standards-based, so much so that it did not require any patented or proprietary technologies to work. That software would be a perfect candidate for open source, right?

Free IT resource TechNet: More ways to know it, share it, and keep it running. Sponsored by Microsoft Free IT resource Attend the SOA Executive Forum: Breaking SOA Bottlenecks SOAExecForum.com/may2007 Sponsored by InfoWorld

That description fits Web browsers perfectly. And yet, despite seemingly having everything in its favor, Opera Software isn't planning to open source its browser code.

"We support open source in that we offer a better browser for Linux," Opera CTO Håkon Wium Lie told me in a recent meeting at the InfoWorld offices, "but we haven't found a business model that allows us to go fully open source."

On first take, that might seem surprising. Browser circles are abuzz with news of Firefox, which is, of course, open source. To compete with Firefox's momentum, you'd think that any other contenders would have to be open source also.

If you look at the actual market, however, it turns out that the majority of Web surfers still use closed source browsers. Microsoft Internet Explorer is of course the leader, but even Apple's Safari , which is built around an open source core, contains proprietary components.

Opera competes in this market, not by playing catch-up with Firefox, but through raw technical merit. For example, Opera has long boasted the best CSS standards compliance of any browser.

"I joined Opera when I saw that Opera was able to implement as much of CSS in three months as Netscape and Microsoft could in three years," Lie says.

In addition, Lie says Opera includes content-filtering options and efficient network access components that make it a popular choice in places such as Russia, where Web surfers typically pay metered access fees for data traffic.

But perhaps Opera's greatest competitive advantage came from some of the earliest decisions in its design. From the very beginning, Opera was built from a very lean and modular code base, particularly when compared with the sprawling Mozilla source tree.

"[Opera] didn't have enough people to make it bloated," Lie quips.

Because of its streamlined code, Opera can build three separate versions of each browser release, including small-footprint builds that Opera licenses to makers of mobile phones and Internet-enabled devices. It is this licensing revenue that makes up much of Opera's current business.

Given that Opera's advantage is purely technical, then, Lie might be right when he says there's no easy way for Opera to transition to pure open source. If Opera gives away the secrets of its advantage by opening its source code, what revenue model is left for it?

The classic open source business model is to give away the software but charge money for support and services around the product. Although that may work for something as complex as a Linux distribution, there probably wouldn't be much demand for support for a product as mundane and familiar as a Web browser.

Other companies, such as MySQL, offer open source software under a dual license. An open source version is made available for casual use; commercial customers are required to purchase a separate license. But relational databases such as MySQL are often integrated into larger products, which can bring them into conflict with the restrictions of an open source license. In contrast, more often than not, a browser is a standalone application. There really wouldn't be much to prevent a mobile phone maker from using the open source version.