Q&A with Mozilla engineering director Chris Hofmann

20

Author: Prakash Advani

Chris Hofmann is director of engineering of the Mozilla Foundation.
For the last eight years, Hofmann worked at Netscape and was involved in every Netscape and Mozilla browser release since Netscape 3.01. Last summer he was hired as the first employee at the Mozilla Foundation and has spent the last few months in startup mode getting the foundation off the ground.

Q: Tabbed browsing, pop-up blocking, and integrated search are some of the distinguishing features of Mozilla. What’s coming next?

A: We are working on developing Mozilla as a platform and Web services are where we are putting the bulk of our efforts.

One of the key strengths of Firebird is that it empowers users everywhere to customize their browser through the use of extensions. Extensions are browser add-ons that are written using our flexible and easy to use XUL toolkit. They enhance the browser functionality in many ways; there are already well over 150 such enhancements at http://extensionroom.mozdev.org with more in development. About four or five seem to show up every week. We’re constantly looking at these offerings for new innovations which we can integrate.

I have a few favorites out of these extensions, but the real value of the extension architecture is that everyone can customize their version of Mozilla in ways that specifically meet their own needs.

Chris Hofmann

I’m amazed that the Google toolbar has so much functionality and is one-quarter the download size of the IE Google toolbar (110k v. 400+k). The highlighting of search terms text in the page makes it easy to scan pages and find stuff. One-button auto detection of language and translation of pages to English is pretty valuable working on a worldwide project such as Mozilla. I frequently use the “translate to English” button to do things like keep tabs on what Daniel Glazman is doing in France with composer work to whatever degree that is possible.

I like the RSSReader Panel’s simplicity of setting the feeds up in bookmarks. Drag an XML feed tag link in your bookmarks, and bang! — you’ve got a new feed. Starting from scratch with a few feeds that I really like also seems better than trying to slog though a bewildering list of feeds that XULchannel and other RSS readers seem to fall into over time. I like the fact that it doesn’t take over my browser to create a different application and gets out of the way when I want to go back to general browsing.

I have one more to add. When I know what book I want, I haven’t found a faster way to find it and buy it than using the Mozilla Amazon Browser. The MAB really shows how useful and fast the combination of gecko Web services and XUL can be to let users perform specific tasks. I’m addicted to golf books and have gotten it down to 20 seconds to find and buy a book from Amazon. For most, it takes at least that long for half the home page to load on the regular Amazon service, let alone the searching, and checkout process.

Q: Reports in the media said that Mozilla 1.5 wasn’t really a significant release. It was more of a bug fix.

A: We attempt to ship updates of all our technology on a regular and incremental schedule to enable others some flexibility in the version that they would like to use and deploy. Each quarter we pump out a milestone.

(Version) 1.4 is being widely deployed by Netscape, IBM, Red Hat, and others. Mozilla 1.5 includes about 1,000 bug fixes and changes over the 1.4 release, plus a couple of new features like spell-checking in mail and composer which many people have been asking for for a while.

1.6 is right around the corner with another set of 1,000 more features, bug fixes, and incremental improvements.

The quarterly release schedule fits with one of the open source principals that guide the project — ship early and often, and listen to user feedback.

Q: With AOL planning to use Internet Explorer and spinning off Mozilla as a foundation, what is the future of Mozilla?

A: The future looks great. More and more companies understand the idea that locking into proprietary solutions has a significant long term cost. We are working with a variety of companies that want to build out key areas of the Mozilla technologies to solve different business application problems. They want to collaborate in the Mozilla code-base and use standards where possible and extend the standards where none exist. We have funding to start coordinating and organizing these efforts, and things are beginning to take off.

Q: Do you have more freedom and resourced than you had with AOL?

A: We are definitely pursuing lots of opportunities. We have been basically operating as a start up company in the last four months getting operations up and running, and looking for opportunities to grow the technology and get it deployed.

Q: Other than AOL, who are the other companies supporting Mozilla?

A: We can’t talk about this much right now other than what was said in the press release when the foundation was formed. As we talk to companies, there is a tremendous amount of interest and support. Getting into budget cycles for major contributions from large organizations takes time and a good deal of effort. There should be some announcements on some additional major contributions soon.

Q: What is the estimated market share for Mozilla and Mozilla-based browsers?

A: We are working with some survey partners to try and round up some good data on this. Adoption and use of Mozilla varies depending on country. In Germany, for example, market share is very high. We are proud of the fact that Mozilla ships in over 99 languages and dialects. The CD distribution program we started with 1.5 is doing well.

There was a good uptick in Mozilla adoption since the release of 1.0 a year and a half ago. Adoption continues to be high, and we attract more new users with each milestone release and additional article in the press. Firebird and Thunderbird are quickly becoming the default clients for many Mozilla followers, and we should see the adoption of those apps take off when we hit 1.0.

More on page 2 …

Q: Can you provide statistics of what percentage of Mozilla users use which platform?

A: Roughly 15% Linux, 7% Mac, and the rest Windows. Mozilla ships in all the major distributions of Linux, and we get a lot of support there. It still amazes me that the code has has been ported to 15 other platforms with the latest distribution the minimal builds running on Compaq iPAQ on the ARM processor.

Q: Galeon implements a nice feature whereby if you shut it when you have open windows, they are saved and are opened when you start Galeon again. Does Mozilla plan to implement something on those lines?

A: I think this is available as an extension on mozdev. I haven’t tried these, but you could give them a whirl …

Session Saver: Remembers loaded tabs and their history items when Firebird is manually closed, then restores the tabs and history items when next started. The saved session can also be manually restored or updated at any later time via the items in the File menu.

There is also another related feature called “Recall.”

Recall: Recall works by tracking your browsing session in a file saved on your local hard drive. When you restart Mozilla after a browser crash (or even an operating system crash), the windows/tabs that were open before the crash will be displayed to you in a pull-down menu that will allow you to return to one or all of the pages you were viewing.

This demonstrates the power of Mozilla as a browser platform, that other developer can pick up work to add features, and you can easily customize your browser to pick up all kinds of interesting capabilities.

Q: Did Apple approach you for Mozilla before they decided to use KHTML (part of Konquerer) for their Safari Browser?

A: Yes, they did an evaluation of Mozilla. Some of our strongest capabilities of the best standards support, HTML editing integrated into layout engine, cross-platform support, and a few other areas weren’t high on their evaluation criteria when we talked to them. It looks like they might be pumping effort into these areas now and retracing some of our footsteps from years ago.

Q: Cellphones, PDAs, and handheld devices are going to be the next big market for browsers. Is Mozilla being used in any of these?

The minimo project is starting to make some ground. We are hoping to see that project ramp up and get more interest in the next year. We talked to many developers at COMDEX Las Vegas 2003, who are using, or plan to use, Mozilla on Linux kiosks and small devices where low cost, no continuing royalties, high reliability, and high compatibility for a wide variety content on the Web are critical. I’ve been running minimo on an iPAQ the last few weeks and really like it for getting news and frequently updated content off the Web. A major cellphone vendor is funding some of this development work.

Q: How many active developers are working on Mozilla?

As the technology matures, and the effects of our investments in automation of the development process improve, we see the ability to do more with fewer people. We had only had a 15% drop in the number of “CVS committers” when AOL discontinued directly supporting the development at the end of July last year. I’m guessing this is surprise to most people.

Number of CVS Checkin Contributors

December 2001 to March 2002: 165

April 2002 to July 2002: 161
August 2002 to November 2002: 152

December 2003 to March 2003: 122
April 2003 to July 2003: 104
August 2003 to November 2003: 88

Beyond the 88 people that committed changes in the last four months, there are hundreds more that attach patches to bugs, and thousands more that download the source code to look and poke at it, and tens of thousands more helping out with testing of nightly and milestone builds. We still have one of the largest and fastest-moving open source project going, with strong development and QA contributions from companies and individuals.

Q: Lots of Web sites are exclusively designed for IE. Is there anything being done to render them perfectly on Mozilla?

We track sites that don’t work in our bug system, and have a volunteer group that evangelizes sites that aren’t adopting standards and are providing proprietary content.

More and more companies see the business case of producing Web standards content to take advantage of speed and maintenance costs. There are a few articles/testimonials on the Netscape devedge site that highlight these benefits and other information.

I’ve seen some studies that show the speed improvements in rendering content that complies with Web standards can be dramatic. When browser layout engine drops out of standards mode into quirks mode it tends to slow down the presentation of information as it tries to figure out all the possible side cases of all the quirks that are possible. Content that conforms to the modern standards can be parsed and displayed much faster.

We really have made a lot of advances in this area in the last couple of years, and we have more support for a wider variety of content.

We look at the nature of incoming bugs, and where we can we provide compatibility to older quirks and proprietary things that Microsoft has implemented.

Q: What is the status of the Mozilla Calendar project?

A: There is a lot of good feedback on the XUL application work that has been done there, but we also see that we need to fill out the back end calendar support to take advantage of the emerging calendar server standards. We are doing some planning now and how to ramp up development work in this area too over the next year.

Q: When do you expect Mozilla 2.0 to be released?

A: We really have our heads down on all of the projects talked about in your previous questions. We haven’t sunk much time into what would constitute a 2.0 release or what work would be involved to get there.

Q: Will Version 2 be a collection of independent applications (Firebird, Thunderbird, etc.) or an Integrated suite?

A: Our preliminary thinking on this is that 2.0 is really a version number for the core gecko components, and a collection of many applications that are built on top of it. Those applications would likely include the latest update to the Integrated Suite, Firebird, Thunderbird, and many more.

Q: Which type of organization prefer Mozilla suite and which ones prefer the individuals applications?

A: It fills a wide variety of needs in many different niches.

We have talked to several enterprise and large organizations that have Mozilla deployments in place, or are considering them. The stable full Mozilla suite provides a good low-cost solution to their needs. These kind of organizations face big costs when trying to do any training, or roll out deployments. So having the full suite around, and not having its UI change much fits a real need with many of these kind of organizations.

Other large organizations want to stay closer to the cutting edge are starting deployments of Firebird and Thunderbird. We think this will ramp up when the *birds hit 1.0.

High costs, security concerns, and the prospect of long delays in the next major release of operating systems gives us a window of opportunity to see the adoption of Mozilla grow over the next couple of years. We plan to fill all the gaps we can with our technology. That’s what is driving decisions about the suite, and the redesigned standalone apps that focus on simplicity, speed, and innovation of the UI.

At the core of this is gecko, which is well-tested technology that renders the vast majority of content on the web and has the best standards support of any browser. All kinds of applications will be built on top of gecko going forward. The full suite was the first major widely deployed app built on gecko, Firebird and Thunderbird will be next, and many will follow.

Q: It makes sense for Mozilla and OpenOffice.org to work together. Both offer suite of applications for the desktop which work on a variety of platform. Is there a possibility of merging the two projects?

A: We are talking to Open Office folks and looking at ways to ensure smooth inter-operation and some integration point between the applications. There are no plans to merge the two projects. I mentioned that we use the same open source spell checker, but that is just one of the areas that we are looking at. There are opportunities to work together on the distribution side as well. The recent announcements of large deployments in Australia and China shows how the combination of these two applications can be combined to provide a powerful solution that meets real needs of large organizations.

I think good integration is what people want. I’m not sure what we would accomplish in trying to merge two large and successful projects.