Just Browsing

Thoughts on XULRunner, Part One: Why XULRunner?

The “Firefox Plus Summit” last week was a wonderful event, bringing together 400 odd members of the Mozilla community, from core Mozilla developers to localizers from around the globe. For me, the key lesson of the summit was the continued high level of interest in Mozilla as a platform. This is a subject that is near and dear to my heart, particularly after the ultimately disappointing rise and fall of Mozpad last year. After numerous conversations at the summit with various concerned parties, I feel I have a much better understanding of why Mozpad failed and why there is still a need to push the platform forward.

It is an undeniable fact that there is considerable frustration in the community over XULRunner. For one thing, developers of XULRunner apps complain a lot about the difficulty they experience getting their patches checked back into the Mozilla repository. This means that they have to continue to maintain a library of patches as they bitrot over time, and it makes them feel like second-class citizens. (Not that this should be construed as a criticism of Mozilla Corporation. Getting a patch into Mozilla is difficult by design, even for those working on Firefox.)

The other source of frustration is that XULRunner is seen as something of an underachiever: a platform with the potential to have a far bigger impact than it has. If Mozilla were to announce that it was stopping all development and support of XULRunner, this would raise a big hullabaloo, but at least their position would be crystal clear. As it stands, XULRunner is 90% of what it needs to be to be a competitive software development platform with many advantages over existing contendors. The missing 10%, along with the seemingly ambivalent attitude of Mozilla Corporation (despite a valiant attempt last year by then CEO Mitchell Baker to clarify the situation), is what causes tempers to flare.

Mozilla’s ambivalence is understandable. The platform has a lot going for it. The vast majority of the code it uses is shared with Firefox anyway, so most of the development effort is highly leveraged.  A number of exciting projects, both commercial (Songbird, TomTom, etc.) and otherwise (Miro, ChatZilla, MobiDVD, etc.), enhance public perception of Mozilla. The Corporation itself is also increasingly using the platform for projects other than Firefox, such as Prism and Fennec. And there are a number of MoCo employees who are strong proponents of platform use by third parties.

On the other hand, Mozilla has a huge hit on its hands with Firefox. Products with such a tremendously large and positive impact on the web don’t come along every day. There is a real risk of the organization trying to take on too much and endangering its strong position in the web browser market, and with it the potential to do much good in the future. When you’re working frantically to release a major new browser version, as was the case with Firefox 3, the last thing you need is to spend cycles on unrelated platform work. Rave reviews for Firefox 3 support the argument that Mozilla should keep its sharp focus on the browser.

There are, nonetheless, at least four good reasons to revisit Mozilla’s decision not “to invest in a pre-packaged or stand-alone XULRunner at this time.” The first is the potential for a strong independent XULRunner to strengthen the Mozilla platform (and thus, indirectly, Firefox). Firefox 3 has ushered in major platform improvements such as enhancement of JavaScript performance, a cleaner threading API and consolidation of functionality into a single “libxul” library. The more individuals, companies and other organizations who are using XULRunner for serious projects, the more momentum can be built up behind future improvements that will benefit Firefox. Some appealing changes, such as moving from result codes to exceptions in C++ and porting portions of the platform to JavaScript 2, would require large engineering investments. The more these changes can be leveraged, the more resources will be available to implement, test and maintain them. This consideration could be decisive in whether such ambitious efforts are even taken on.

Secondly, a more widely used XULRunner will broaden the base of developers with Mozilla development skills. This will, in turn, increase the pool of developers capable of creating quality extensions for Firefox. Considering that the extension ecosystem is arguably Firefox’s biggest asset, this could be of considerable importance to the browser’s ongoing success and market share growth.

Thirdly, a strong platform commitment will provide Mozilla with some measure of future-proofing. The annals of business history are littered with companies who foundered because they were hesitant to endanger existing product lines. (IBM’s relunctance to jeopardize its mainframe and typewriter businesses with a strong move into the microcomputer market is a classic example.) Firefox is king right now, but who knows what the future will bring? Already we are seeing new browser directions like Prism and Fennec that rely on Mozilla-as-a-platform (and are experiencing some pain when the platform functionality is not as generic as it could be). By commiting to a strong independent XULRunner, Mozilla will be better prepared to take appropriate action as the browser landscape evolves.

Finally, there is a clear market need for an open development framework for multiplatform internet-enabled applications. Microsoft is pursuing this market with .NET and Silverlight, but the company is hardly known for its openness, and it will always favor Windows even if some support is provided for other operating systems. Adobe’s offering is also proprietary and crucial components are not open source. As this space heats up, many of the applications that would traditionally be deployed in the web browser will be implemented instead as so-called Rich Internet Applications that run directly on the desktop. Turning XULRunner into a truly competitive offering would be of great benefit to the web at large, in the same way that Firefox has brought more vigor and openness to the browser space. It would also be a great way to promote the principles outlined in the Mozilla Manifesto.

It is worth nothing that, by all accounts, Firefox was not embraced immediately by much of the Mozilla community when it was first released. While practically everyone recognized an increased focus on user experience as a Good Thing, many objected to the sacrifices that this entailed (such as discontinuing support of the Suite, with its integrated mail/news client, HTML editor, chat client, etc.). If too conservative a stance had been taken, Firefox might never have been released and its subsequent success would have never materialized. This isn’t to say that all change is good or all risks worthwhile, but any long-term success is predicated on taking at least some big scary (but calculated) bets.

In part two, I’ll discuss what steps are needed to help XULRunner achieve its full potential as a competitive RIA platform.