Firefox Mobile Threatens Mobile App Stores, Says Mozilla

Barence writes "Mozilla claims that its new Firefox Mobile browser could be the beginning of the end for the hugely popular app stores created by Apple and its ilk. Mozilla claims Firefox Mobile will have the fastest Javascript engine of any mobile browser, and that will allow developers to write apps once for the web, instead of multiple versions for the different mobile platforms. 'As developers get more frustrated with quality assurance, the amount of handsets they have to buy, whether their security updates will get past the iPhone approval process ... I think they'll move to the web,' Mozilla's mobile VP, Jay Sullivan, told PC Pro. 'In the interim period, apps will be very successful. Over time, the web will win because it always does.'"

Comments

Seems Unlikely

[saisuman]

I saw my wife playing Assassin's Creed on the iPhone today. I can't imagine a game of that quality being remade in Javascript unless it comes with some funky O3D-like capabilities.

Two things.

[Hurricane78]

Two things, my friend:

1. Java is THE dominant platform if you want to program anything that works on pretty much all mobile phones on the planet. Apart from the iPhone, and some Windows Mobile phones, I don’t think there is a phone that can’t do Java.
2. In the real world, not many people care about the App Store or the iPhone. It has only 3-4% percent of the global market share, and technologically already was surpassed when it came to the market in Japan, was a novelty for about a month in most of Europe, and only in the USA has gained more than 10% for obvious reasons. Which means, others are still hugely dominant. So much in fact, that I don’t even think it’s worth targeting the iPhone platform. (I’m sorry, but if you now think I’m trolling, that’s the reality distortion bubble, created by the hype. I’m in no way hating the iPhone or anything. It has great raw power and a good UI. I’m just stating the facts as I know them from actually being in the market, and keeping up to date, because I need that to make a living. Prejudice is just stupid, and am happy to be corrected. :)

So I really see no point in yet another layer of inner-platform failure, to use JavaScript, when you already have fast Java with accelerated OpenGL, EAX-like audio support, and tons of functions. (Be aware that as much of it is accelerated, Java on mobile phones is vastly faster per raw CPU power, than on desktop VMs.)

If they can offer me all those hardware-accelerated APIs, an ability to check if the phone supports them, a fast JavaScript compiler, and 96% of all phones of the world having it pre-installed, I might consider writing for their platform. ;)

PastryKit

[fandingo]

Ars Technica had an article about a hidden framework that Apple was developing before Apps hit with 2.0. http://arstechnica.com/apple/news/2009/12/pastrykit-best-iphone-web-app-library-you-never-heard-about.ars

Actually looks pretty cool and could allow more web-based apps.

I still think that local apps will be preferreable. The thing is that a lot of apps are only useful on the web, so the concerns about not being able to access them w/o a net connection are baseless. Not all apps, but there's lots of social networking apps and others that need networks.

Re:web-app-web

[dgatwood]

Local storage is pretty cool, and the CSS animation and stuff doubly so. There's still the problem of having a good way to back up local storage.... *sigh*

That said, it's still not even slightly close to what you can do in a native app. Even if performance was identical, the DOM is beyond half assed as a GUI environment. There's not even a drag and drop mechanism built in that works across all browsers. There's no way to guarantee that your handlers won't get stripped out by some overzealous UI library that you load. Basic functionality like contentEditable (for WYSIWYG editing) is barely supported in any browser, replete with hundreds of serious bugs that make it very hard to deal with. There's no way to set up an automatically recurring callback with a guaranteed period. There's no way to spawn multiple threads of concurrent execution (except for a FireFox-specific mechanism). There's no standard way to talk to hardware. And those are just the huge problems.

Even simple things like specifying which UI elements should grow proportional to the window size is an utter pain. Creating clickable buttons that don't get their text content selected can be rather entertaining. Convincing the browser to not deselect the selected text in a contentEditable region when you do so is doubly so. Then, you have that fun box model that only a committee could love (all of us are dumber than any of us). Don't get me started on trying to do column layouts with CSS. I could go into specifics, but if you've ever tried to build any significant web application, you're already nodding in agreement.... :-)

Yeah, it's going in the right direction. It's got a long way to go, unfortunately. Right now, it takes mounds of custom GUI libraries just to get usable UI, mainly working around the fact that the web browser just wasn't designed to do this stuff. When I can write a web app that's lightweight and doesn't require bringing in something as heavyweight as Prototype just to get anything done, we'll be at least in the right ballpark./p>