Adobe, Apple, and Flash | Hacker News Reader

Adobe, Apple, and Flash

daringfireball.net

197 pointssophiebits100 comments

With the recent Vimeo and Youtube announcements that Flash would effectively become a fallback delivery format in the future, people have been too caught up in the Theora discussions to say much about this potential deathblow for Flash.

As Gruber says, when people say Flash support, they basically mean Youtube. Once Youtube goes native, all the other sites need to rethink their usage because Flash is no longer necessary on phones, tablets etc. and indeed, even for platforms that support it I can see Flash blockers and selective Flash usage becoming popular for battery life reasons alone.

Flash is still the system of choice for any video requiring a low level of protection, e.g. iPlayer, hulu and needing near-universal coverage.

> As Gruber says, when people say Flash support, they basically mean Youtube.

He doesn't say this at all. He says "Flash is the way video is delivered over the web". Video is only delivered by Flash, but that doesn't imply that Flash only delivers video.

Games are a huge source of Flash usage on the web. 75M people played Farmville last month alone. That's not an insignificant audience. Gruber points this out in the article, along with the reason why Apple doesn't care about Flash games.

I don't disagree that smaller sites will probably take the lead of larger sites regarding front-end technology switches. But I believe this follow-after will happen more because HTML5 is the "cool new thing" not because Flash has somehow overnight become obsolete.

The problem with Flash is that it's really, really slow. For most web-browsing PCs there's more than enough power under the hood to spare to account for the slowness of Flash (even then, Flash is sluggish on most machines, just not enough to be really annoying).

On a phone or other slow-ish device, I cannot imagine how it could even hope to run.

Look at the Palm Pre - a very good example of trying to run non-native code all over the place in a mobile device, and the whole thing grinds to a halt. Multi-tasking on the Pre is a bit of a joke because of it - sure you can do it, but not if you wanted some semblance of snappiness in your UI.

Sorry if I was unclear. I'm not arguing that mobile devices should run Flash. To be honest, I don't think they should, for the same reasons as you. I merely clarified that Flash has major uses beyond video and for that reason the adoption of HTML5 isn't going to kill Flash.

"The problem with Flash is that it's really, really slow."

if you program a simple particle generator with say a 400x300 display, flash is a TON faster than canvas\javascript.

Which browser?

Can you post numbers by any chance (I'm genuinely interested - in my experience chrome and safari were sometimes several times faster than ff with canvas)?

No, Flash itself isn't slow, but a lot of people are making a lot of Flash content that is hideously non-optimized, and a combination of that on a webpage can make anyone's computer slow to a crawl.

And a precious few people are capable of creating good Flash content that is optimized and does run fast. Look at this for example:

http://www.rockonflash.com/demos/pv3d/macworld/demo/as3/inde...

Adobe's excuse about not having a hardware acceleration API is a red herring, and a painfully obvious lie if you think about it for more than 5 seconds. There are several reasons I have come to this conclusion.

1. This is clearly demonstrated when you play a video "natively" through flash on OS X. Watch as the < SD resolution video immediately hogs an entire core of your CPU; in fact the lagging starts almost from the moment the flash object is initialized! Even a trivial operation like clicking fullscreen on a YouTube video locks up the entire host process and beachballs Safari for 5+ seconds as it sits there with it's thumbs stuck up it's arse, causing you to miss whatever was playing at the time.

2. Ok, so now try playing the same video through XP etc on VMWare or Parallels on the same machine. What's this? It's perfectly responsive and uses a small fraction of the CPU power? Apparently it is ≈5x less CPU intensive to emulate and deal with the overhead of virtualising an entire PC, it's CPU and associated operating system than it is to run the same thing natively on the host OS. *Note that due to limitations inherent in virtualisation, there is no access to any kind of video acceleration whatsoever, let alone assisted h.264 decoding. You have a framebuffer and that's about it (and even that highly abstracted, double buffered, composited and managed by Quartz as with every other window on the system)

3. My original MacBook which I still use has no hardware assist (GMA950) - yet QuickTime manages playback of fullscreen or windowed 1920x1080 h264 at around the same CPU usage playing a fucking YouTube video. 4. Flash video on the mac has always been excruciatingly slow, even before h.264. I remember seeing iMac G3 600Mhz's struggling to maintain even 15fps with postage-stamp sized h.261 (which is what YouTube used to use) just 5 odd years ago. A G3 600 is several orders of magnitude more powerful than what is required to playback video of that complexity. A G4 867 can do 640x480 h264. Clearly something is very very rotten in flash, and getting more offensively pungent by the day. Which now brings me to:

What in the everloving FUCK could their code possibly be doing?!?!

I think their programmers have quite clearly gone insane and subscribed to (or invented?) the Rube Goldberg school of software engineering - I don't see any other rational explanation.

I'm quite certain we could have cured cancer while making SETI look like the equivalent of someone with a funnel stuck in their ear and pointed at the sky, what with the aggregate difference in CPU cycles of mac flash vs windows.

Meanwhile, if Adobe can't even manage to lie effectively about the reasons why their own software is so utterly broken, I'm not holding my breath for a resolution. The first step to fixing a problem is admitting you have one, though if you ask them it's all Apple's fault.

To think that people want flash on the iPhone. God almighty, haven't we all suffered enough?

Flash games only exist because everyone installed Flash to watch video.

Before flash video hit, the installed-base wasn't that impressive/threatening.

Precisely the reverse is true. People installed Flash to play games and use artist / restaurant 'websites' made by new-media-douchebags.

Flash video took off because of the enormous install base, and the ability to control presentation.

> artist / restaurant 'websites' made by new-media-douchebags.

This is hilarious.

I find this hard to believe.

A google answers query from 2005 puts Flash penetration at 98.2% (http://answers.google.com/answers/threadview/id/561724.html)

The article introducing the groundbreaking sIFR library from 2004 cites the reasoning for using Flash as high penetration (http://www.mikeindustries.com/blog/archive/2004/08/sifr).

Macromedia in the year 2000 puts their penetration rate at ~90% (http://web.archive.org/web/20001109224100/http://www.macrome...)

There is simply no evidence that points to video causing Flash to become a standard plugin.

Flash was included by default in Windows XP. It also has a simple update process.

Video was barely emerging back when Flash began, ala RealPlayer and the like. Flash itself didn't even begin getting good video functionality till the 6th version (Flash MX) from memory.

Unfortunately, youtube is a big reason for needing flash but not the only one. If a company wanted to build a website in the past and probably even now, they could easily be convinced to spend a lot of money building a flash based site. There is a lot of legacy flash type things that will not be replaced soon.

I don't even think it's just people spending a lot of money building a site. Outside the tech sector Flash is the #1 choice of every graphics designer turned wannabe web designer. These guys generally have no idea how to code but can sell to small business customers with a few canned flash effects. So you see a lot of flash in smaller sites too.

No APIs for hardware h.264 decoding? What, you haven't seen OpenCL? You can't write a GPU-accelerated video player? Mplayer's been GPU accelerated in h.264 (well, x.264) for a relatively long time, on both Linux and OSX. Are you saying a behemoth like Adobe can't learn a trick or two from someone else?

There's no excuse for Flash to be so bad.

edit: for a better, in-browser example, tried Unity lately? In-browser GPU accelerated 3D graphics. And it's never crashed any of my systems, in any browser. ( personal favorite: http://blurst.com/ )

edit again: hah, check this out. Note the "helloracer.com" link in the top article, and that Unity works on the iPhone. http://drawlogic.com/category/unity3d/

I did'nt know about this, but +1 to you. It would be interesting if this could be brought to Gruber and/or Adobe's attention.

I think you can't access them as a browser plugin though.

It's also a total canard: hardware decoding is virtually brand-new in Flash. I think it only came with 10.1, right? But the performance problem and disparities have been around for much, much longer.

With things like version 9, Adobe were much more upfront. There was a puff blog touting how 9 was much faster on the Mac because they were no longer doing some ludicrously expensive font-checking in the middle of (what should have been) very tight loops.

Which do you buy: Adobe can't make Flash reasonable on the Mac because of mean ol' Apple, or they choose not to because of (understandable, given how Apple treats them on things like Carbon) institutional bias towards Windows and a nasty, hairy codebase?

Sorry, but OpenCL is not API for hardware H.264 decoding as we know it on other platforms. Yes, OpenCL allows you to run code on GPU, but that's where it ends. It is no match to DXVA in Windows or VA-API/VDPAU in Linux, which allows you using the silicon dedicated to decoding MPEG-2/4 ASP/4 AVC/VC-1. There is no equivalent to these APIs in OSX.

Which still leaves the ability to speed up h.264 / video in general with GPU rendering, which is far more optimized for it than your CPU. It's not ideal, but it's significantly better than nothing.

Granted, it's not "true" h.264 hardware decoding, but it's better than pure-CPU-through-Flash, and as far as I can tell they have yet to take advantage of it, and have been exceptionally late to the game no matter what.

That's difficult to say without benchmarking. Yes, you can offload some computations to GPU (you certainly are not going to parse the bitstream or VLD decode using shader programs), but do you get net gain, when you are shuffling buffers over PCI-E? You also limit your target audience to NVidia 9400 and up when talking about OSX, basically telling the Macbook users with Intel graphic cards, that they do not need to apply (Intel 3100 and 4500 are perfectly capable of decoding H.264 and VC-1, just not through shaders).

It's easy to see for yourself how many crashes are caused by Flash. Firefox used to crash regularly for me. After I installed FlashBlock it can run pretty much indefinitely.

I agree. I don't mind Flash when it's used for something useful, but most of the time (perhaps even 99.99% of the time) it's being used for ads. Those ads are, by definition, very attention grabbing because of their use of animation. The animation uses my CPU.

So, in general I say... go away Flash and browse without it enabled. It's very, very rare that I need to access a web site's Flash.

Seriously, what do you think would happen if, magically, all those Flash ads disappeared one day?

Bloated, gif-based animated ads and/or HTML5 <canvas><audio><video> animated ads would take their place. (With, in many instances, much worse performance).

Flash isn't the issue; CPU-pegging animation, and ads, are the issue. The Medium != the Message.

what do you think would happen if, magically, all those Flash ads disappeared one day?

Users of FlashBlock, ClickToFlash, and platforms that don't support Flash (such as the iPhone) already know what would happen, and they like it.

No, they know what happens if the ads disappear for them, but if they disappeared for everyone, they would just be replaced with other sources of annoyance and bother. Yes, you could block some of those, but eventually, you start blocking things that are genuinely useful, like images or javascript. Right now, Flash one most sites means one of two things: ads or video. Kill Flash and now the situation won't be nearly as clear or easy to block.

Unless you've got a solution for the underlying problem that it is profitable to distract people, this is going to happen no matter what people do. Removing Flash from the list of ways which they are able to do that means that they have to work that much harder, rather than you having to work that much harder to conserve your attention by blocking it. Make the bastards chase you.

That's just it, though...blocking flash is easy and 99% of the time, all I'm missing is ads and video. And one click displays the video if I really want to watch it. Whereas if flash goes away, I'll have to start blocking things that are more useful on more levels, like images, javascript, etc.

To be fair Flash doesn't help the issue.

I make games in Flash and performance is frequently an issue I and others battle against. The point at which Flash performance begins degrading is stupidly early.

The solution in that case is to stop browsing websites that use annoying intrusive ads.

It's not just poor sites that use annoying and intrusive ads. Many newspapers and otherwise high quality sites have intrusive ads. It isn't easy to simply drop all of those sites and remain informed.

The Mac implementation of the Flash player is awful, it regularly pegs my CPU. I'm assuming that HTML 5 in a well coded browser will not do the same, no matter how annoying the ad.

If you're talking Youtube's HTML5 video player, I regularly stay below 20% CPU. That's 10% total power, as percentages go above 100 when using more than one core.

Otherwise, it'd be Javascript + <canvas>, which so far is pretty darned fast, but not too friendly to ad-servers. Not Flash-fast, but speedy, and quickly getting faster. If JS+canvas gets more popular, you can bet that JS will get WAY more attention in this area.

I'm actually quite disapointed by Youtube's HTML5. I had great expectations, I know it's just the beginning and I'm sure it'll be better, but I'm closer to 80% CPU than 20% with my Macbook on Safari 4.

That could explain it. I'm on a MBP, with 10.6, so I may be getting the hardware decoding, while you're not.

On Ubuntu 9.10(all same video):

Youtube HTML5 - 15-20% CPU

Youtube Flash - 20-25% CPU

.flv in Totem - 2-3% CPU

Clearly, there is optimisation work needed here.

Any kind of JavaScript while(true) loop will lock up the page, so you aren't safe from incompetent coders just yet. Depending on the browser and the OS that can bog down the CPU as badly as Flash.

(And there's no move towards parallel execution of in-page dom-touching JavaScript yet - web workers can't do it. Concurrency is such an issue that you can't even pass a function into a web worker in case it has bound variables - it's separate-script-and-message-passing only.)

Seriously, if Flash disappeared, ads made by canvas or SVG inserted via obfuscated javascript is harder to block than plain Flash ads.

Mozilla's crash stats often show Flash as a very large percentage of crashes. You can see a dozen or so crash lines in there that happen in NPSWF32 or "Flash Plugin":

http://crash-stats.mozilla.com/query/query?product=Firefox&#...

I think that plugin instability is a big problem overall, probably just as big as the 32/64-bit plugin divide, which itself was a serious blocking problem on 64-bit Linux for a while.

I'm on 64bit Linux here, and as long as I used the Adobe Flash plugin, it regularly crashed (in various ways, most of the time just the flash subsystem).

I know there is a 64bit version now, but I got genuinely fed up with them, so I deleted the plugin and use swfdec for youtube and just don't care about the rest. Can't say that I'm bugged much, my browsing experience actually improved.

The main reason I boycott flash is still, that it is by no means sustainable and it inhibits the progress of the web a lot (get the feeling of a hamster in a hamster wheel, solving the same damn problems over and over again). It must simply die.

I used to think these claims were a bit bogus. However, my macbook pro was locking up solid and completely unresponsive at seemingly random times requiring a hard-reset (running 10.5.x). The constant thread I noticed was that I was always using either NetNewsWire or Safari when it happened. I kept trying different things to fix the issue (erasing preferences, verifying permissions, etc...) but no luck at all. There was nothing in the crash logs and no kernel panic or anything like that for me to even trace the culprit. It would just completely hang without so much as a wimper.

The last thing I tried was that I installed click-to-flash. I haven't seen one of these crashes since. I thought for sure there was no possible way it could be flash since the mac usually does a pretty decent job of managing hung apps without bringing the whole system down, but for whatever reason it really looks like flash is causing this. I suspect it's some odd problem with my flash installation or configuration.

I've had the same problems with my machine. With both 10.5 and 10.6. Flash just sucks on a mac.

It shouldn't be possible for it to lock up the computer, though. It doesn't install drivers. No matter how you slice it, that is Apple's fault.

I hear this a lot, but it must be a thing for Mac users mainly. I've never had flash crash Firefox on the PC, and I'm a flash developer a lot of the time. Interesting.

Flash causes more crashes than any other application for me on both OSX (Safari) and Windows (Firefox). It's horrendous. Flashblock / ClickToFlash did the same for me, I nearly don't crash anything any more, and I don't fear losing all my tabs by opening a random website.

As to the "flash where flash is useful", hopefully JavaScript + the canvas will eliminate the majority of those cases. There are still some cases where flash is significantly better, but they're pretty rare. Most of the time I highly suspect it's popular because of the nice editing tools, not because the producer(s) wanted to target Flash.

Realistically Flash isn't going to just disappear. I think everyone understands that. We just want an alternative whenever possible. I can certainly understand that Flash game creators don't want to re-write their perfectly functional games in Objective-C. The Adobe Flash iPhone complier solves that problem nicely. For anything on the web, ads, videos, navigation, etc I don't think it's unreasonable to expect a Flash free alternative. Adobe's embarrassing Flash support on OSX/Linux is a perfect example of why we need alternatives. Adobe apparently cannot even adequately support three major operating systems. How can we rely on one company to provide make or break support for an entire assortment of computing devices? Three major operating systems (at minimum) each with 32bit/64bit variants and multiple browsers in mainstream usage. Half a dozen mobile phone platforms too. To date Adobe has been able to offer good Flash support for a grand total of one operating system. If there weren't real fundamental problems with the security, stability and performance of Flash I don't think we'd be having this conversation at all.

"If there weren't real fundamental problems with the security, stability and performance of Flash I don't think we'd be having this conversation at all."

But that always seems to be the way one technology gives way to another, because it improves on various aspects that were lacking in its predecessor.

I think an apt comparison is RealMedia versus RTSP audio streaming. RealMedia had a significant market share (for the time), but the mishandling of the product led to people finding/building/demanding better tech.

It's arguable whether Adobe is mishandling Flash at the same level, but Flash could easily go away in favor of HTML5 and Javascript.

The only thing lacking right now is the tools, and the Flash IDE would not be difficult to improve upon. On top of that, there's already precedent in the Palm Web OS IDE. Make a similar kind of interface for building easily deployable standards-based web applications, and Flash could end up as a bad dream faster than you think.

I completely agree with the judgement that Flash is the cause of most application crashes on Safari. It is time for flash to die. Not only are most flash plugins terribly inefficient (demanding 100% processor power on both cores to run a simple movie is ridiculous) but they also tend to consume more memory than they should.

The only people who would want flash to stick around would be flash based ad providers and flash based games. Movies and other formats are easily dealt with using HTML 5.

I think it would be best if Apple created their own flash implementation, yet Adobe probably wouldn't want them doing that.

The Apple way to play H.264 is through the QuickTime APIs (and really, as of Snow Leopard the new QuickTime X APIs), not to write your own H.264 playback code that seeks to directly access hardware accelerators.

I'm confused... If hardware decoding were simply a matter of using the QuickTime APIs, why doesn't Flash just use those?

Using QuickTime: "Decode this video for me." Using hardware directly: "Do this math for me."

Even if it were technically possible for Flash (as a plugin) to call out to QuickTime this way (and it may well not be), it is undesirable from Adobe's perspective to depend on code they don't control for something so central.

It would be very easy for flash to use QTKit to just play a video, in the HTML5 kind of way. But it would be quite a bit harder for them to do it in a way that remains compatible with all the different play/pause/volume/seek controls that different video sites have built in to their flash apps, not to mention the advertising schemes. If they weren't smart about how they handle all the compositing, they could probably end up really slowing things down when there are things like custom translucent controls overlayed above the video.

With total control over the iPhone OS, I wonder if Apple will slowly start repositioning OS X as an operating system for power users and focus more on the iPhone OS for it's consumer-line products. A basic iLife/iWork suite for iPhone OS would be all that's needed for most users.

John Martellaro had an interesting point on twitter:

> MacBook = content creation tool. iTablet = content consumption tool. I doubt we'll see any writer's apps. No terminal app, of course.

http://twitter.com/jmartellaro/status/8041600445

I think that's dead-on.

That's exactly what I expect to happen. Steve said the tablet is the most important thing he's ever done; I believe that's because he ultimately intends it to replace general purpose computers for most consumers. If it's as locked down as the iPhone, I fervently hope that does not come to pass.

Buy Adobe. Owning Flash source code problem solved.

Market cap of adobe as of writing is 17.82 billion. It's been rumored for a while and might make a lot of sense (or none at all).

It's much cheaper to push Flash further into irrelevance by growing their new platform(s) without it. The things that Flash can do that just straight up can't be done with HTML/CSS/JS is actually pretty small, and competitive pressure is likely to push JS performance ahead of ActiveScript in time, if it isn't there already.

Noooooo! Adobe is at least as much of a monolithic-program-overlord as Microsoft is! Arguably worse, given its updating application that intertwines with your computers' innards. It's practically opposite of what OSX is encouraging, with fast, simple applications everywhere that talk to each other.

They could buy Adobe and just drop the entire company into a dumpster along with it's software.

Even better, buy the rights to h.264 first. License it freely for all online uses. Scoop up the scraps of Adobe for a song.

There's a lot of h.264 to buy, from a lot of owners, and not very many of them would dream of selling.

Besides, Apple has been a member of MPEG-LA for a long time, and has 1 patent in the h.264 pool: http://www.google.com/patents/about?id=VNKRAAAAEBAJ&dq=7...

Indeed, they could write that check twice without debt financing.

Daniel Eran Dilger of RoughlyDrafted has a very cogent perspective on most things Apple.

Also, has Gruber ever linked to/acknowledged Hacker News before?

"Were Flash Lite to gain momentum, it might make Adobe the Microsoft of mobiles, and Flash Lite the new Windows. That also makes it obvious why Apple wants to choke Flash to death before it falls into position as the new lowest common denominator in proprietary platforms on a new crop of mobile devices...

And you thought the iPhone was just Apple's way of muscling into the mobile business! No, it’s really a proactive battle against a wide swatch of proprietary platforms promising to plague a new wave of mobile devices."

http://www.roughlydrafted.com/RD/RDM.Tech.Q2.07/879DD82D-559...

I think the whole 'open web' argument goes down as soon as Gruber mentions h.264 as the alternative to flash. That's replacing an evil with another evil.

I think h.264 is a bit less evil than flash. First, it is not ubiquitous yet. Second, it turns the flash vs <video> fight into a <video h.264> vs <video Theora> fight. This move already lean towards open standards and free software. Pushing further can only mean the eventual victory of Theora.

Now, talking like Theora doesn't exist is evil. However, Apple and its fans don't have much choice: Apple is a licensor of some patents on h.264, and hardware acceleration isn't available for Theora in the Iphone. A world without Theora is in their best interests.

The eventual format will probably be determined by Google (with YouTube). Will they chose a non-encumbered format, or will they choose to support the Iphones ? I don't know.

I agree that it is a lesser evil right now, but it is not an that's why I think Gruber's "open web" argument is hypocrite at least.

when I see how much flash content of the web is useless advertising and I combine that knowledge with the currently available limited traffic plans available here in Switzerland, I think flash on the iphone would only work if I can selectively chose which flash applet I want to see which is probably too complicated for apples taste.

Even worse: While I'm sure it's possible to create optimized flash-content that is not too taxing on the hardware, I'm not quite sure that the various ad agencies take their time to optimize their output for small devices. Why should they?

So flash-ads would probably really diminish the overall browsing experience - just look at the flash-enabled android phones out there: The moment you hit a page with flash ads on them, they slow down to a crawl.

Would you want to have that experience on your iphone?

After two years of heavy iPhone use, I might (direly) miss background applications or real over-the-air podcast download but I certainly don't miss the flash support. At all.

I'll be a witness to the fact that memory management on the Flash platform is horrible. The virtual machine that runs ActionScript is pretty good, but any use of the Flex libraries, "debugging" versions of SWFs, programmers who don't understand low-level concepts, etc... quickly leads to memory bloat, and in extreme cases - frequent crashing.

Try loading in SWFs created with Blaze (the artist tool) into your flex app, and you're in for an even more huge world of hurt. You can fume memory (crash the browser within 30 seconds) without even writing a single line of code, just by having an artist placing certain kinds of keyframed animations and fonts. We spent weeks trying to explain this problem to the absolutely idiotic Adobe Support Gold, and after months we got something like "seems like something might be up. we'll have our engineers look into it. Can you send us another test case?"

Tablet not supporting Flash would imply that the tablet would also be tied to the app store? That's me losing interest right there.

I'd expect a tablet to be more like a "real" computer than a phone. Where (i)phones would be real computers, too, except they are tied to app stores, which limits their use as real computers.

It being a (hypothetical) new product Apple would imply +appstore, -flash. I wouldn't expect anything new from Apple would allow unsigned binary installation as a default.

It sounds like you should probably just tune out anything new from Apple and save yourself the time.

Apple still makes "normal" PCs (MacBooks, iMacs, Mac Pros), which all allow unsigned binary installation as a default. Possibly the tablet could fall into the PC category rather than the "mobile gadget" category?

It sounds like you should probably just tune out anything new from Apple and save yourself the time.

Sadly I'm getting to that point. I love my MacBook Pro and Mac minis, but I left the iPhone for Android, and am expecting the tablet to be another piece of excellent engineering deliberately crippled by Apple's marketroids.

Anecdote: A few years back I wrote a Flash 8 piece for a relatively large ad campaign. The code ran fine on everything except OSX/Safari. Not only did it run poorly, but it had the strangest behavior: it ran poorly the first time you visited the site with Safari, but if you reloaded the site, everything ran fine.

I bought a Mac mini to try to debug the problem, but without dev tools I was at a loss. After sharing my frustration amongst my circle of friends, a guy that did a lot of OSX work decided to take a look with some debugging tools.

From what I recall, he discovered that Flash was making a perfectly valid call to an API method that should have been deprecated. The bug report was filed with both companies and I can only hope that one of them paid attention. The hack to get it to work on OSX/Safari wasn't pretty.

I would like to have an issue clarified: The author mentions that flash is completely controlled by adobe which makes apple uncomfortable. This is true practically, but theoretically, the swf format is open. Apple could come out with their own player if they wanted. Am I missing something important?

Here is the spec for flash player 10. http://www.adobe.com/devnet/swf/pdf/swf_file_format_spec_v10...

There are a lot of open source flash projects. Here is a nice list I found:

http://osflash.org/open_source_flash_projects

The amount of effort to implement a new Flash player is huge. Epic, even. Considering how fast Adobe is adding new features to Flash, trying to clone it would be like getting on a treadmill whose speed is controlled by your competitor.

Yet there are non-Sun JVMs...

Next to video, the largest pro-Flash argument is games. Some have said that most flash-based games could also be made using nothing more than what is available in HTML5.

If this is true, is the reason these games are built using Flash just a matter of the quality of tools available to game developers? If there were an open-source equivalent to Adobe's Flash developer tools, who's output was HTML5, would Flash developers switch?

I think it is a major reason. I'm currently designing a project that I intend to use flash as the frontend for. Its only a desktop app, but the flex builder tools seem so easy to work with that I've decided to make it web based. Open to alternative suggestions.

There are a bunch of reasons we use Flash for games

- we give a site a SWF and that's it, they have the game. HTML5+CSS+JS+images+sounds+music+a flash bridge to handle audio, etc is a mess by comparison to "a swf".

- distribution is critical to the business of making Flash games, when you see a game on whateverwebsite.com that begins with the logo and branding of for instance www.armorgames.com, that's because Armor has paid the developer to brand that game with his site. This alone is a massive industry that includes many amongst the most-trafficked sites on the met.

- power and functionality, every generation of the software Flash gets noticeably better. Massive progress every other year, that allows developers to do things like use webcams for player movement, audio and visual sound effects, graphical work, only some of which right now is still years from being viable across the board in HTML5/CSS/JS. New versions also enjoy rapid adoption rates by ocnsumers.

- truly cross platform, even if it sucks a bit on linux or whatever, we as developers don't have to do anything at all to accommodate anyone at all ... we write once and that's it. The fantasty with html/css/js etc standards making this a reality for that platform have historically and currently not worked that well, the most obvious example being video codecs for html5, where you'll now have to sniff the user's browser and serve them alternate content.

Looks like I'm not the only one thinking about this:

http://www.zeldman.com/2010/02/01/flash-ipad-standards/

In particular:

"Developers who supplement Flash with HTML5 may soon tire of Flash - but Adobe has a brief but golden opportunity to create the tools with which rich HTML5 content is created. Let’s see if they figure that out."

Is it easier to create Flash packager for iPhone( http://www.adobe.com/devnet/logged_in/abansod_iphone.html ) or 64bit Flash Player?

Flash wont die out anytime soon.

You wont find any platform that is so well equipped for rich experiences as the Flash Platform. HTML5 is no where near that.

A lot of developers hate flash, but the rest of the world loves it.

okay, what's the current best web solution for vector animation besides Flash?

Canvas. Or SVG. Those are the iPhone-friendly solutions.

For IE compatibility, experiment with Raphael: http://raphaeljs.com/

Are you suggesting artists or non-programmers to create animations not by using an authoring tool, but by typing numerous for's and setTimeout's ?

Also is it possible to streaming the animations in SVG?

I think that if canvas and SVG gain some momentum, then authoring tools will follow. The biggest trouble is getting the momentum though.

I am thinking the opposite. If HTML5 can draw the attention of more non-programming crowd, like gamers, artists, stylists, etc. it might have a shot.

Currently the HTML5 canvas is just, well, a plain canvas.

Are you suggesting artists or non-programmers to create animations not by using an authoring tool, but by typing numerous for's and setTimeout's ?

1. Write new authoring tool that outputs numerous fors and setTimouts

2. ...

3. Profit!

That's exactly what I'm thinking as I read this whole debate. Does such a tool exist, and if not, why not?

Does someone want to sell a cliff notes version of that?

I think this article highlights how backwards Apple can be on things. There are really no advantages to having your browser be 64-bit (I could imagine how you could have a 4GB+ browser process but its really not the most common use case).

I have really been impressed with how well Windows does 64-bit. The 32-bit layer just works.

Also, Apple is always a pain about anything GPU related. Usually you have to buy through them at an absurd premium, then you are usually dealing with an older opengl version. They are not the platform to use if you are interested in having decent graphics acceleration.

You're wrong. 64-bit Intel means more performance as the cpu has more registers, and 64-bit apps on Mac OS X runs in a more secure runtime environment.

>64-bit Intel means more performance as the cpu has more registers

Sometimes. First of all its a bummer Apple only sells Intel. AMD makes quality chips. 64-bit means better performance sometimes. Usually you take a hit because of larger pointers. LLVM can do pointer compression, it'd be interesting to know if Apple uses it widely. 64-bit really is awesome when you want to use lots of memory. Otherwise 32-bit is perfectly fine.

I've been a 64-bit enthusiast for quite sometime and I can pull up plenty of Gentoo posts from back in the day advocating the wonder of registers. Not all algorithms magically improve in performance when they are available unfortunately.

Also, no mention of my GPU argument? I'm sure you won't say there is more performance to be had there.

I'm sorry, you've got that totally backwards. The way Windows handles 64-bit is decidedly disappointing: it's a whole separate version (no upgrades!) that has two copies of everything.

Apple has pulled it off beautifully: one version of the whole OS, Universal Binaries for everything, uses whatever kernel is most appropriate because it can run 64-bit processes on a 32-bit kernel, everything just works.

I don't think I do. There is a certain irony in you saying everything just works, when the blog post explicitly says that hardware acceleration doesn't work for flash, and seemingly by Apple's design.

You hate there being a whole separate version of the OS, far out. I think the needing universal binaries is a little silly. The fact is Windows biggest strength is its gigantic amount of software available for it and it doesn't require rebuilds all the time. Stability can be a good thing.

I find Apple constantly releasing upgrades a little ridiculous. It's nice to know my PC will be able to run just about anything without need to spend $100 every 2 years.

It's nice to know my PC will be able to run just about anything without need to spend $100 every 2 years.

Snow Leopard didn't cost anywhere near $100. I paid about $30 for it.

That's Stasis, not Stability. Microsoft hasn't introduced a popular new native API in a decade!

Reply on news.ycombinator.com