Decoding Adobe's Big Device Push
nerdyH writes "Adobe yesterday chummed the waters around Flash and AIR as cross-platform app dev environments for mobile devices. It promised runtimes for several popular mobile OSes, including WinMo, Symbian, Palm webOS, and Android, with future RIM/Blackberry support hinted as well. Moreover, it reiterated its commitment to the Open Screen Project, an Adobe-led industry group that, if you deconstruct its name and look at its membership roster, appears tactically focused on enabling hardware acceleration of Flash/AIR on devices, as part of a larger strategy of making the runtimes ubiquitous as UI development frameworks for essentially every computer-like device with a user interface."
Woot I win!
they should totally use arm
With HTML5's video, audio and canvas elements, there will be less and less need for Flash in the future on the web. It seems like Adobe is realizing this as well and has decided to move the focus of Flash from mere embedded objects on web pages to a way of easily creating full, rich and cross-platform applications for both PC's and phones.
This coiuld work out pretty well for them in the end. I must admit clicking a game together using Flash and publishing it to every major platform sounds more attractive than the more traditional ways of developing software, and I'm sure I'm not the only one who's thinking this.
Pretty good is actually pretty bad.
...they will learn something from squeezing Flash onto these embedded devices that can be used to help make the desktop edition less resource intensive.
Adobe yesterday chummed the waters around Flash and AIR as cross-platform app dev environments for mobile devices.
Literally?
Noooo. It's bad enough that Flash slows down and eats system resources in Windows, Mac's and Linux, now they want to inflict the same on underpowered mobile devices. That's sick!
Take Nobody's Word For It.
Oh, I'm sorry, I though this was an argument. I'm in abuse. Terribly sorry about that.
I can escape from that damn room on every device I own!
What can you do with Flash that you can't do with html5?
I think you should perhaps go back and read the summary carefully.
You're the first one to mention open source. Unless you think the story submitter spells source in a really eccentric fashion.
I don't expect morality, equality, consistency, or justice from the law. I expect only legality.
Certain projects shouldn't fork. Sun wouldn't open up Java for the longest time, because they didn't want forks of Java, and they didn't want to repeat what they went through with Microsoft.
I propose a new license that operates on a few basic principles.
1 - You can redistribute and modify the source code.
2 - You may compile the original source code, and even compile modified versions for personal use.
3 - You may not redistribute modified binaries.
In this scenario, users can compile themselves, test, fix bugs, write patches, etc. They can submit patches upstream, but upstream still largely controls the project and prevents major forks. You would still attract community developers.
I think a license like this would work well for Flash.
http://blindscribblings.com - Tasty pop-culture in conceptual fashion.
Sometimes a maintainer refuses to give up a project, but refuses to continue meaningful development. Consider the X.org fork.
I could be wrong, but..
Unity3d.com is probably doing what Adobe plans to already, except they're using .NET. Cross-compiling code into real iPhone applications. I haven't dug too deeply into how Unity3d is doing it, but it seems pretty clear -- you can write your code in .NET with some pseudo-alternative languages like 'Boo' (python), and it makes you a nice iPhone binary that'll pass Apple's deployment criteria.
Considering Adobe has the time, money, and smarts to do it, don't be surprised when their 'Program Actionscript for the Iphone!' system is a very tightly defined API coupled with the iPhone framework that is cross-compiling..
Before I gave up on Perl, the assertion that Parrot would be some fancy answer to everyone's programming problems by allowing you to program in any language you wanted. I somewhat scoffed at the idea, but more recently as I've been working with ARM processors and doing a lot of cross-compiling work I can understand why it's an important idea that will soon be second nature to us.
If I could buy stock in Unity3D right now I would, because those guys nailed it. They just need to scale up and out of just the 3d game market.
In this day and age, anywhere you could use media rich applications, you have a web browser.
Presumably, if Adobe doesn't establish Flash as a cross-platform dev environment for mobiles, then Microsoft will manage to foist Silverlight as it's own bloated slow lane for mobile devices. And the same devs that give us IE-only web apps will start producing Silverlight-only stuff for mobiles.
Now maybe Miguel would disagree, but I think it's better to have a truly cross-platform bloated enviroment than to have a single-platform bloated environment (I assume Silverlight/Mono is at least close to Flash in bloat). Sure, I'd take streamlined before bloat, but cross-platform trumps streamlined.
By the way, aren't Android apps based on Java? Since when is that a paragon of efficiency? Or does Google use some kind of 'compiled to machine code' Java variant? Likewise WebOS apps - aren't they largely Javascript? Who said mobile device platforms weren't bloated already?
Posted from my Android phone. Oh, I can change this? There, that's better...
Except for the Adobe part.
And that Flash thing.
Forgive me if I don't trust a company that can't write a plug-in that will give me less than 80% CPU usage (480p) on my brand new Macbook Pro. The Linux and Windows version are also glacially slow, and resource hogs. Frankly I want less Flash, not more. If Adobe can't get their shit together on the 2nd largest OS platform, how the hell are they going to get it working well on a teeny mobile ARM core?
Having hardware video acceleration seems to be one thing, but hardware acceleration of an API or bytecode seems like it could lead to chip bloat, which would make chips much more expensive to make. Shouldn't we concentrate on optimizing the software side of things? Would increasing the processor speed and battery capacity be cheaper than specialized silicon?
I must admit clicking a game together using Flash and publishing it to every major platform sounds more attractive
Slap it together and call it a day!
Never mind it doesn't take advantage of platform specific features. I'm sure users wouldn't care about THAT at all. I'm sure your sales will be just fine...
Sometimes easier things are just easier, not better.
"There is more worth loving than we have strength to love." - Brian Jay Stanley
...and Adobe claimed they would have flash on Android this Fall.
October is here. Now they say next year.
I am not hopeful that they can get flash on Android. Possibly they are waiting for better devices so they don't have to shoehorn it into the G1, which could use more RAM, but it is what it is.
In fact, I predict, no Flash for the G1 ever. And many of the other platforms as well. Adoby wants to FUD the developers and keep HTML5 on the shelf as long as possible, since stuff like Canvas will pretty much eat their lunch and dinner if they don't watch out.
deleting the extra space after periods so i can stay relevant, yeah.
So this means I will finally be able to use flash sites on my G1 and HTC Touch? Sweet
"Computers are a lot like Air Conditioners" "They both work great until you start opening Windows"
Adobe is saying that the new 10.1 version will fully offload h264 decoding to the GPU. If that works as advertised, then it would solve lot of problems involving full screen video playback on websites like hulu and youtube.
HTML5 is kindof awesome, but even the most awesome technology is limited by the number of people who can use it. Unless the W3C or Microsoft or Google or the Mozilla Foundation manage to convince the world to upgrade their browsers with the speed that Adobe can upgrade the install base of the Flash Player, HTML5 is always going to play second fiddle.
Now according to Adobe, Flash Player 10 is at 94% adoption in mature markets, and that's about... what, 10-12 months after release? The HTML 5 spec was formally named in January 2008, and the original started in 2004. Admittedly- corporate IT departments (the big evil) are as unlikely to upgrade the Flash Player as they are the browser, but if it takes that long for anything to make it into HTML, Adobe will have already had several upgrade cycles to react, improve, and move on.
Having said that: we can always return to the days of browser specific web sites, and that'd force people to upgrade: "This website is optimized for [Insert favorite browser here], please change your browser".
This signature can save you $400 on your car insurance!
Flash's real strength is on the content-creation side, and the fact that most Flash is generated by "designers" not "developers". All the HTML5 specs in the world won't displace Flash if they require a team of Javascript/SVG gurus to use. There needs to be designer software on the same level as Flash, and that's not a trivial problem.
So by "designers", do you mean: people who don't care about not breaking standard GUI controls (buttons, scroll bars), who don't care about breaking the "back" button and bookmarks and the ability to trade URIs, who don't care about their site not being indexed by search engines so people can find it? Those people?
While a useful tool to make certain sites more "interactive", I agree with Jakob Nielsen when he says that its use is 99% bad. As a UI building system, I have never seen a page that needed to be in Flash. Currently the only reason to have it is video streaming, and hopefully that will end with HTML5.
This article claims that the new Flash player they're working on for mobile devices will only support ActionScript 3.0, meaning the majority of flash apps which are http://www.appleinsider.com/articles/09/10/06/html5_assault_on_adobe_flash_heats_up_with_clicktoflash.html&page=3
That really sucks for anyone who trusted Adobe and stuck with their Flash Lite development guidelines, all previously made mobile apps have just been made useless.
So really if you want Flash on your mobile you're only going to get stuff that's just been made with you in mind. So there's basically few ActionScript 3.0 apps out there and people are going to be going "I thought I had flash on my phone, why won't it play all those games I like?"
Now for those websites with older Flash games and apps, they either redo them in AS 3.0 or just move to HTML5 and support everyone. Upgrading to AS 3.0 had better be REAL EASY or they've just shot themselves in the foot here.
Cwm, fjord-bank glyphs vext quiz
Really, I've yet to hear a serious answer to what you can do with flash that you can't in modern javascript.
Clearly they have pulled developers from the Flash-optimization-team-for-non-windows as well as the optimize-size-of-acrobat teams to get this to work.
People have been complaining about the lack of 64-bit Flash for four years now. If Adobe develops this plan with just as much passion as they had for x64, it'll be 10 years before we have to worry about it.
Canvas is just svg all over again. Remember when SVG was going to kill animations in flash because it was *technically* possible to achieve the same thing? Canvas is no different. Having a technical solution is not the same thing as a workable solution.
That's all I want from Adobe. Please please please please please!
I love Mondays. On a Monday, anything is possible.
The reason SVG isn't/wasn't a workable solution is that IE doesn't support it. That's not an issue on mobile devices.
Let's say you're work for a small business or non-profit and you have, say, four licensed copies of Adobe's Creative Suite products. When you bought them, the current version was CS62. Several months later, you've decided there's value in providing these tools to another employee. That's right, you want to give Adobe another $500-2500, depending on the product(s) included. Sorry, Adobe is now at CS63, and they won't sell you CS62 anymore. What's more, the two versions are "sort of" compatible. You can import one version into the other, but all elements may not translate properly. Nope, if you want to add another user of their software, you have to purchase another full license and four upgrades in order to keep them all at the same version!
Did you get a lot of life out of your Photoshop installation but finally decide to upgrade? Great! Check online for eligible upgrade versions. Hooray, it's listed! But wait, when you attempt to install it, it won't accept the license key from your old version of Photoshop. It turns out Adobe's installation path for your version is to call customer service. [insert dead horse about how ridiculous it is to punish your paying customers vs. pirates by forcing them to activate] So you get Adobe on the phone. "We're sorry, in order to upgrade your version, you're going to have to uninstall the product, then reinstall it again with a special command line switch."
Personally, I will avoid Adobe products wherever there is a viable alternative. Adobe chooses to follow the Microsoft example of exploiting dominance in a sector by putting their customers through bullshit those with a choice would never put up with.
Has anyone else tried to use Flash/Flex to do anything more then just a game/animation? The frameworks are at best immature. While the Flex language is "ok" (basically just ECMAScript), the libraries, tools, frameworks, and most everything else that goes with it are just abysmally bad when compared to any other modern language (Java, Obj-C, Python, C#, etc).
The Flex Builder plugin for eclipse only supports Eclipse 3.3, which means modern plugins won't work. The flex compiler itself is slow and hard to setup. Oh, and the tools only officially support Windows and OSX. The documents are horrible and only give you most simplest of use case examples, but this might be because most of the frameworks breakdown when doing even remotely off the "rail" they have defined.
Just as a quick example of something inexcusably broken, the ComboBox that comes with Flex doesn't have a set selected by value function. You can only set by index and by label... which is just crazy when you consider most ComboBoxes contain localized strings order alphabetically in that local.
As a development environments go, I think you would be hard pressed to find something worse for development than flex.
And I could go on about how bad the user interaction is by default. But you really have to see that one to believe it.
If not by HTML5, or DHTML (4) + JS, etc... then by the forces that went after Microsoft. I mean, if you can declare Microsoft a monopoly - even though no one ever held a gun to my head and forced me to use Windows or Office - then Adobe is dead with Flash and the way they throw that crap around. "doesnt work on the iPhone because of technical issues...." BULLSHIT. Its ALL about licensing and the Adobe CEO likes telling stories.
Well, get on with it!
Why bother with browsers at all then? They might as well come out and say everyone should just ditch free web browsers and just install an Adobe plug-ins or a Silverlight runtime, buy their application servers and IDEs, and code in the language that they dictate instead of more open and continually developing web standards that the majority can agree on. That's what Adobe and MSFT hopes. They still think it's 1999 and they can take over your web experience and tell you what is "Rich Internet Experience" and what is not. The only 2 reason Flash continue to be used are 1) annoying ads, and 2) porn. I still haven't met anyone who thinks an entire site developed in Flash provides an enjoyable experience to users after about 2 minutes besides Flash developers and managers that hired them.
I don't know what I am anymore and I've done UI in both Flash and other systems. I don't think Flash is very efficient, that's the problem I have with it. And there are a lot of developers mucking with Flash code that probably should be doing something else:)
I think therefore I can't be ~TTNH
h264 decoding is not the problem. In fact, 80% CPU usage when playing back a simple video is probably the -best- case scenario for Flash on Mac, since there's not very much Flash actually going on. Try some games from armorgames.com and compare your framerates to a similarly spec'd Windows box. On my 2006 MBP, many of those games are simply unplayable due to slide-show level framerates (10fps). The CPU, of course, is pegged at 100% no matter what the application.
I have to install Flash blockers in my browser because sometimes I like to surf the web while on battery, hahaha. Flash alone cuts my battery life in half.
Face it, Mac Flash performance is like every other poorly, hastily executed Mac port: shit poor. I saw an Adobe developer claim on someone's forums they knew what the problem was, but just didn't have the "resources" or corporate will to fix it.
Erm it's not hinted at it's been announced.
I was amazed when I saw current Flash player (10) could play "futuresplash" the original flash files all fine even with added hardware acceleration.
To have such backwards compatibility without adding bulk to a plugin which is in version 10 is the true secret why web designers love flash.
Here is the futuresplash demo from 1995 http://www.4dm.com/files/tech/blue.htm
Believe or not, some execs at Nokia still believe one may upgrade his hardware to have a fixed/better feature.
The entire scheme of things in Symbian is to have current software/features and even beta experiments on latest generation of hardware (devices) regardless of which capabilities the older generation may have.
Basically, it means E71, the number 1 suprise hit of Nokia won't get the attention of N97/5800 which entire Nokia seems to focus on. They actually believe they can convince a E71 owner who is perfectly happy with his phone to buy a N97 to have "better software". Even Apple can't dare to do such thing, they still support first gen iPhones with current software!
For example, I got a E65 which is plain Symbian S60 V3 without "feature packs" (hw/sw mixed extensions such as FPU). It has horrible, I mean horrible issue as "app manager" (add/remove programs in winland) takes almost a min to appear. They won't fix it as there is "E66" in same series now. Nor I won't have "password saving" web browser.
Before I write more on this strange issue which people outside Symbian land may have problem understanding since it is a bit insane... If this "Flash open screen", the "desktop flash on mobile" only ships on Nokia 5800 and N97, it is Adobe's millions which were wasted and it is no way to race with iPhone in major scheme of things. It is not "Symbian", it is basically latest models of Nokia. There is no way they should dare to talk about 100M devices, it is cumulative amount of Symbian shipments.
Not on OS X it won't.
Besides, even *simple* flash is a dog on OS X.
Let me tell the reason you get 80% CPU usage while doing a simple(!) thing as playing a 480P SD embedded file.
Apple didn't provide application/codec developers any way to access GPU features existing since 2004 as "hardware h264 decoding". They even couldn't access them themselves until Snow Leopard. In Snow Leopard, perhaps Flash player in current major version or minor update, separate tree will have 10% CPU usage but NOT the earlier OS X versions. Also if there is a XCode forced issue like "build for 10.6 or 10.5 but not for 10.4.11" like the one happened to VLC guys, they won't even touch it, it will be a seperate tree to download for snow leopard which will be a real support nightmare.
Of course, they could make video decoding with CPU way more lighter but it would add bulk to the plugin and some amazing amount of development time would be needed just because Apple developers were lazy to code driver h264 decoding interface.
Your GPU which is perfectly capable of decoding the H264 content without touching CPU stays idle and does just "2D acceleration" which is like 1% of its true power. It is not just Adobe to blame if we talk about video decoding/playback. In theory, Apple could ship 10.5 upgraded drivers and support Developers with hardware accelerated decoding right now. Why wouldn't they? Sell 10.6 and new hardware. That is what keeps OS X half of what it could be right now.
I bet all my powerpc macs saying it will be only possible with 10.6.x and not pre 10.6.
Only Snow Leopard offers the interface of doing that thing and of course, it won't stop people from blaming Adobe when it ships :)
Those little games you talk about nears a billion dollar industry which actually saved Real Networks from going down the drain with Real Arcade.
Funny thing is trying to explain it to a slashdot poster as you do. They think poor content provider/game developer morons are sticking with Flash because they have no clue about "standard mpeg" or "svg".
BTW I know some PS3 owners who installed Linux/PS3+Gnash+IBM Java just to play those tiny Flash/Java games on big screen.
Will the mobile device vendors except Apple who is a complete genius to have a single line of model and a method of easily falling back of features ship "updated browser" to their line of 3 years of different devices? They never do.
Vendors like Opera, who makes their money with browsers on mobile can do such a huge task but in return, they would ask for money. Try to explain to Nokia execs that their "webkit" browser is a disgrace and worst possible way of using webkit framework on any OS/device and they should go back licensing Opera. I wonder how many minutes it will pass before security comes to escort you out of building :)
Eh, the current Linux version does that already - "fully offloads h264 decoding to the GPU" - via OpenGL. It's just that it's the wrong way to do it due to all the layers it's got to go through to get there. It sounds nice, but I suspect that the article is just marketing buzz.
If 10.1 offered any significant improvements on the desktop, there'd be useful beta builds not only available right now but also in wide use by the geek community.
~/ssh slashdot.org ssh: connect to host slashdot.org port 22: too many beers
ship an application on time
I am not following you. Mobile browsers like opera should be following the same standards as all of the other browsers right? Don't the common web standards require html5 compatibility?
Just as a quick example of something inexcusably broken, the ComboBox that comes with Flex doesn't have a set selected by value function. You can only set by index and by label... which is just crazy when you consider most ComboBoxes contain localized strings order alphabetically in that local.
What, you mean like this? (took me 3 minutes to put this together)
<?xml version="1.0" encoding="utf-8"?>
<mx:Application xmlns:mx="http://www.adobe.com/2006/mxml" layout="vertical">
<mx:Script>
<![CDATA[
public function doSomething():void {
for(var i:int=0;i<testComb.dataProvider.length;i++){
if (testComb.dataProvider[i].data == ti1.text) {
testComb.selectedIndex = i;
}
}
}
]]>
</mx:Script>
<mx:ArrayCollection id="ac">
<mx:Array>
<mx:Object label="A" data="1"/>
<mx:Object label="B" data="2"/>
<mx:Object label="C" data="3"/>
<mx:Object label="D" data="4"/>
<mx:Object label="E" data="5"/>
<mx:Object label="F" data="6"/>
<mx:Object label="G" data="7"/>
</mx:Array>
</mx:ArrayCollection>
<mx:TextInput id="ti1"/>
<mx:Button id="b1" label="Selected By Data" click="doSomething()"/>
<mx:ComboBox id="testComb"
dataProvider="{ac}">
</mx:ComboBox>
</mx:Application>
It was originally produced by Macromedia.
We'd need either a fork or a complete rewrite to support all the existing content.