Ten years ago next month, I moved to California to work on Mozilla full time and one of my first discoveries was a metal worker building giant metal dinosaur sculptures. I suggested to folks at Mozilla, back then, that we get one. Today, thanks to the efforts of Tiffney Mortensen, we finally have one.
He (She?) isn't quite as big as the ones I first saw, but that makes the t-rex office friendly and that's a compromise I'm happy to make.
If you're on Firefox, you have to check this out. Very cool! Ignoring the numbers makes it very difficult but super awesome!
This report from AMD seems to suggest that IE 9 will indeed support the </canvas> HTML element. If that's not a slip up, that's really quite awesome. Go team IE!!!
I told everyone tha tIE9 was gonna rock. Get a preview.
CSS3 selectors, <video> and other HTML5, SVG, border-radius, perf perf perf, addEventListener, rgba opacity, and more.
Welcome back, Microsoft!
There's a bit of confusion about what's happening with the Mozilla JavaScript engine so I'm thrilled to be able to point you to David Mandelin's blog post, JägerMonkey & Nitro Components, (and re-point you to Chris Blizzards post) to hopefully clear things up a bit.
I'm not particularly technical, but here's what this all looks like to me.
A couple of years ago, all JavaScript engines were pretty slow. Then Apple introduced SquirrelFish Extreme, which optimized large swaths of JavaScript quite well using something called a Method JIT, and Mozilla introduced TraceMonkey which optimized fewer cases but to a very high degree using something called Tracing. Then V8 from Google came along with something similar to what SquirrelFish Extreme (now called Nitro) was doing.
The end result of this first round was that Firefox was really-really hyper-fast at a few things and behind somewhat on many other things. Now, with JagerMonkey, Mozilla is adding a Method JIT like Nitro and V8 have so we'll be well optimized for a majority of cases while preserving the Tracing approach where we can be super-duper fast when it's applicable.
To build this method JIT, we're re-using a relatively small part of the Apple Nitro JavaScript engine while building a bunch of new functionality ourselves.
But there's even more cool news. Once we get the method JIT in and we're on solid footing with all of the other JS engines for those common cases, there's still a lot of opportunity to improve our Tracing feature to cover more and more cases. So hopefully we'll see increases in the places where we can be out of this world fast, while making sure there are no places that are slow. That's enough to make you wonder if Apple or Google (or Opera) will decide they need Tracing in their JS engines. If they do, I hope they take a look at Mozilla's code so they can benefit from our work.
Following in Apple's unfortunate footsteps, Microsoft has just acknowledged that it too will restrict all installs on Windows Phone exclusively to a Microsoft managed store. That's a real shame.
This quick review of the 7 tier two browsers on the European browser ballot is pretty much spot on, except the part where they give somewhat positive marks to two of the browsers on the list.
Google to strip unique client ID from future Google Chrome installs.
But why ship it at all. Is it really that important to track individual users through their first automatic (and silent) update?
I just read Taras' great blog post Extensions & Startup and it got me thinking that we should just automate this and advertise it with extension listings at addons.mozilla.org.
All you'd need to do is have some automation that gets triggered when an AMO editor approves an extension that measures its impact on startup and publishes that as a required part of the extension's description in the gallery.
That way users could evaluate that against their desire for fast Firefox launch time and make informed decisions about which add-ons they're going to use.
This might also inspire extension authors to consider focusing more on ameliorating start-up impact through various techniques like delayed loading.
While I don't think Microsoft has fully opened the valve, we are starting to see some downloads from the European browser choice program.
Unlike one of the other browser vendors on the ballot, we're not seeing a tripling of downloads. Then again, I wouldn't expect that given our baseline of 2M downloads per day.
We did just cross the 100M Firefox 3.6 downloads. Those are all manual downloads because we haven't turned on the "prompted update" system yet. When you're getting a couple million downloads a day, the tens or even hundreds of thousands of additional daily downloads coming from the browser choice screen don't really stand out.
Still, I hope that when Microsoft finally opens this to all of Europe (which I don't believe they have, contrary to recent press accounts) I think we'll start to be able to see more meaningful numbers.
I'm thrilled to be able to report that the new release of Opera, 10.5 (Windows only for now) has pretty decent Open Video support.
I did run into several problems, like scrubbing back and forth being prone to video errors and not handling application/ogg, but overall I'm very pleased to see another Ogg(Theora+Vorbis) implementation released.
After I've used it for a few weeks I'll try to report back on any other findings.
