Looking Ahead to Firefox 3.6: Speed Matters

Apparently, one billion served isn’t enough for the Mozilla Project. The project announced the release of Firefox 3.6 alpha 1 this week, code-named “Namoroka.” This release starts the purge for native platform widgets, JavaScript speed improvements, and support for new Cascading Style Sheet 3 (CSS3) properties.

This blog discusses the push to get rid of native — i.e., platform-specific — widgets in the Gecko layout engine in Firefox. The long and short of it is that Firefox should be drawing elements the same across platforms, scrolling should be more uniform, and so on. For Firefox and Web developers, this will matter. For end users, it shouldn’t matter much except that they’ll see a more uniform experience if they’re using Firefox across platforms.

If you have 3.6 handy, you can see all kinds of pretty gradients on the Developer section of the Mozilla site. If you view them in 3.5 or Google Chrome, you’ll just see a bunch of white boxes. This is because the gradient feature in 3.6 is proposed only, and specific to Mozilla for now.

This release also includes a proposed feature to specify background size, and multiple background images. All in all, some nifty stuff, but nothing Earth shattering.

Is it Faster?

Since there’s not a lot of new features or glitz in 3.6 as of yet, I focused mostly on the performance. The Mozilla Project measured the performance of recent Firefox versions using the SunSpider JavaScript Benchmark, so I tried that with 3.6a1, 3.5, and Google Chrome.

Google Chrome scored 704.0ms, Firefox 3.6a1 scored 1730.2ms and Firefox 3.5 scored 2971.8ms when running the suite.

So, is 3.6a1 faster than 3.5? It looks like it is. However, it still has some improvements to catch up with Chrome.

Of course, this wasn’t the world’s best-controlled test. I was running the test on an openSUSE 11.1 system using a Lenovo T61 ThinkPad with 3GB of RAM and an Intel Core 2 Duo clocked at 2.4 GHz. Results might vary on different OSes and different hardware configurations. And it might be a good idea to run dozens or hundreds of times rather than a handful of times to get the most accurate results. And, of course, we all know that benchmarks aren’t perfect representations of real-world performance anyway — but they at least provide some rule of thumb to measure by.

Want to see how your favorite browser fares? Try it here.

Interested in testing Firefox 3.6? I recommend not using your existing profile — or backing it up before you begin. To start a new profile, run firefox -ProfileManager -no-remote. This should ensure that if Firefox won’t eat your profile if it happens to have any major bugs.

I’ve been using Firefox a little bit less lately, but I’m glad to see that the Moz project is focusing on performance improvements in 3.6. Prior to using Chrome, Firefox was always fast enough for me — but the combination of stability and speed in Chrome have been very alluring. But I do miss some Firefox features and extensions, so I’m keeping a close eye on Firefox 3.6 development.

The schedule for 3.6 is pretty short. The first alpha was released just a few days ago, and the final release is set for November — so less than six months after the Firefox 3.5 release was final. We’ll take another look at 3.6 development as it gets closer to the final release in November, when the UI pieces start showing up in the builds and all new features are in place.

One thing that’s not in this release, or planned at all for 3.6, is process separation. Process separation is the feature that allows each tab to run in its own process, so if one site causes a crash, it doesn’t bring down the entire browser. (Google Chrome has this feature.) I’d like to see this one sooner rather than later, but it’s not a trivial feature to implement.

It’s a bit of a shame that Google opted for building its own browser rather than helping the Mozilla team to implement this and other performance improvements — but in the long run, users may win out as the two teams keep trying to one-up each other feature and performance-wise.

If you’re a hardcore Web developer or dedicated Firefox tester, then be sure to check out Firefox 3.6 alpha 1. It’s pretty stable from the time I spent using it, so the “alpha” label is probably scarier than the actual software. But, there’s nothing compelling in the release yet to check out for most users.