Gone in a Flash: Adobe's long march to HTML5

Never owned an Apple product, but...

I still have to say "THANK YOU, Steve Jobs!" for giving Flash the boot.

Finally!

Where as I sincerely appreciate what Flash has down for the internet, I have always been a sceptic for its use.

In my experience it slowed down machines, crashed websites, and needed a further update/plugin/install for it to work. The same went for Silverlight.

I prefer nice, clean code that all parties, browsers, and devices can use with little or no additional fuss.

I'm suprised at Adobe to be honest. most companies would flog that dead horse (note: its not a dead horse just yet) into the ground but it seems they are looking to the future with a more unified web platform. Scary stuff.

hello...goodbye

As a successful professional web developer, it's one technology I've never learnt to code and now I'll never need to.

Brilliant!

The Flash Orbituary

Flash was a technology which supported the creation of polished-looking browser applications of questionable utility. Some websites used it to annoy visitors with mad-hopping graphics flashing over the screen.

On of the stronger aspects of Flash was a proper video codec, which made it popular to view videos of exploding wales and copulating humans.

Most infamously, Flash was a major security risk on every computer equipped with it and many severe network-based attacks exploited the shoddy implementation quality of this Adobe product.

The death of Flash is well-deserved and will relieve the global computer used community of a dangerous piece of software. Flash shall burn in eternal hell.

Bit of history being rewritten

1. Apple didn't like Flash because it ran like a piece of crap on the Mac.

2. Adobe failed to address Flash performance on the Mac over several years so Apple rightly were cautious about the iPhone.

3. Apple are totally correct when they talk about phones having limited power and limited battery. Do we forget the old days when a Nokia 6300 could do 10 days+ on standby? The fact that Symbian was designed to be low power (low memory and low cpu overhead), or the fact that the iPhone only lasted a day because Apple underestimated how much people would play with their phones?

And on the back of that, Adobe couldn't be trusted to develop a low power version of Flash.

What killed Flash was iOS and Adobe's intransigent. No iOS, no future.

However, what I never understand is why people got so upset in the Apple vs Flash thing, calling Flash the open standard when it was a £700 package you had to buy if you were going to develop on it. It wasn't an open standard in the slightest and had nothing to do with the internet, which is largely based on open standards.

BTW, how is Adobe doing with a long overdue rewrite of Photoshop or whatever they call it these days. A hugely expensive product that has been neglected for 10 years as they dodge the development budget for OSX.

To all the flash critics

I am not a web programmer. Well, I wasn't. C++ and Delphi, SQL and an odd assembly instruction or a perl one-liner here and there were my tools of choice. Only when faced with porting a C++-written project over to Web did I first consider using Flash. That was around 2006, and I've never felt sorry about that choice. Compared to HTML/JavaScript/css it is everything they are not: strong-typed, predictable, performance-optimized, and, above all, uniform across platforms. I can run the same application on Windows, Linux, MacOS, Android, WebOS and Blackberry, and even compile it for iOS, without changing a line of code. And it looks the same, be it the oh-why-still-used IE6 or the newest Firefox. Javascript and CSS has literally to be written thrice, for different browser flavors, and the design is mostly dictated by browser quirks instead of proper architecture concerns. When you waste more time on workarounds than on actual implementation, it means the platform is wrong. Standard - compliant browsers? They couldn't manage it for 15 years, who says they'll manage it now in HTML5? Not to mention that no HTML5 technology replaces flash capabilities when it comes to graphical objects and customized interfaces. "canvas" can't do vectors, and SVG can't work in all browsers yet, despite decade of development. And DOM is slow. VERY slow.

In short, I don't see alternatives for Flash in the field of rich internet applications. The shortcomings mentioned above (can't copy, can't scroll, etc) are the fault of the programmers, not of the platform. They are "holding it wrong". To see, how it looks when done right, see Photoshop online, for example.

As to security faults... There were (and still are) enough Javascript and CSS and TTF and Jpeg exploits in the past. With canvas, sound and video tags we will see brand new exploits. Flash is not a major source of those, and is only there in the limelight due to its ubiquity - no single browser now reaches 50% of the user base, and flash has 99%.

I'll get my coat, yes. The one with Flex SDK cd in the pocket.

mixed feelings

I was a Director developer, who tried FutureSplash Animator in 1996 and couldn't take it very seriously. I regarded it as a toy. We had afterburner and then shockwave, which could do so much more. Director's scripting language was never highly regarded, but had powerful LISP-like features, a command-line, acceptable OOP, and it compiled to bytecode. You could do great things with it. (Most of the really great Director stuff was overlooked, or made for very small audiences). The early 'action' editors in Flash were a bad joke.

I was amazed to see the younger upstart 'toy' technology bought by Macromedia, and then steadily capture the attention of the producers of cool but gimmicky content, banner-ads, and ultimately most of the development budget of Macromedia. It was as late as Flash 5 that a proper scripting language was introduced - a javascript dialect now known as ActionScript 1. Verity Stob wrote some hilarious comments about this period of ActionScripting on the Reg. I advise the keen reader to check out the relevant article.

I was even more amazed that a handful of people started using it to make some quite decent casual games, even emulators, synthesisers and some fabulous data visualisation tools (e.g. gapminder). Slowly, steadily, it became more technically powerful, and certainly more up-to-date than its older brother.

I switched to Flash, even began teaching it, and developed a couple of solid medium-size applications with it, plus many small things. I adopted AS3 and grew to like it, but there has never been any doubt that Flash - authoring and playback - has been rotten on Apple's systems for almost 10 years. Crashes, hangs, lousy resource management (memory/cpu) and poor OS-integration have been the norm. Multimedia designers - many of whom are Mac users - have always had a love-hate relationship with Flash, and I believe that now they are ready to move on.

I confess to a certain amount of schadenfreude. I remember the snooty Flash kids, their tool of choice in the ascendant, looking down their noses at Director devs, just as we looked down our noses at the Hypercard and Authorware community. It becomes increasingly obvious: Closed multimedia authoring systems are always a dead-end, no matter how defacto 'standard' they may temporarily be. Microsoft's never-popular Silverlight - a potential competitor crippled by neurotic strategy - is another relevant example. In each previous case, there was always an obvious proprietary ship to jump to, but that is not so now, and that is why Flash SWF will linger on, way past its sell-by date.

So instead multimedia designers are expected to bank on HTML5 and javaScript. In theory, a great idea. In particular we are offered Canvas, and a fairly ropey, underdocumented audio/video playback API, with various ideological and technical encumberances. (Did you know that iOS can only play one piece of HTML5 audio at a time? Or that Firefox doesn't and will never support mp3?)

Now I see an interesting dilemma for multimedia designers and content/front-end developers, who cut their OOP teeth on Java, ActionScript or C#: A mindset migration from strict-typed classical languages, to javaScript - a language which superficially resembles Java/C#/ActionScript, but in truth is more like an exotic variant of LISP. How will the multimedia designers cope with this? How will they adopt javaScript when 99% of javaScript books and web-based tutorials promote dilettante or sub-optimal practice? How will the multimedia design courses teach it? The paradigm shift is going to be interesting to watch, and can only be ugly.

Then there is code editing. Flash's ActionScript editor is not the best editor in the world, but it has some very friendly features which really help you learn. It is, quite frankly, ideal for first-time coders with artistic leanings. Is there a similarly accessible javaScript code editor that offers syntax-checking, automatic code-indenting, friendly error feedback etc.?

Is there a browser which doesn't just give up silently when your javaScript has the tiniest error? Does the browser error console offer much help? The error feedback in Flash authoring has never been amazing, but at least it gives you a clue where to look for problems in your code.

I know that this is total non-problem for comp.sci folks, but the audience for the Flash authoring tool is quite different. They have different needs and different expectations when making interactive stuff. And yes, many of them are crap, and clueless but many really want to make good stuff and adopt best practices.

The landscape for content developers, self-taught game devs, arty-nerds, interaction designers, ux/ui designers etc. for creating javaScript-based content is fragmented and unfocused. Hardcore coders will always be happy with Ultra-edit, or VisualStudio, or Eclipse or notepad++, but those tools will never catch the hearts of the folks that come from an arts or design background.

I also see no other software which has vector drawing tools as friendly and intuitive as those found in Flash since its very first versions, and I see no animation tools that can export lightweight vector-based animations (e.g. svg) for the web. (BTW The animated gif exporter of Flash *really* sucks, and animated gifs are surely not where we want to go in the 21st century).

Also, canvas offers no sprite model, no 'movieclips', no collision detection, ultimately you just have a bitmap which does not distinguish one mouse or touch event from any other, except in pixel coordinates. If you want to keep track of individual visual objects, you have to build your own engine in javaScript, which means performance overheads. There are a few of these engines out there, some are very good, but it's a terrible shame that some of the features that multimedia designers regard as basic or fundamental ("I want something to happen when I click on this moving monster") are simply not available, straight out of the HTML-5 toolset.

SWF is undoubtedly on its deathbed, but the Flash authoring tool has no obvious heir. The market is wide open for a "Flash-killer" which offers some, or all of the features mentioned above, which generates HTML5+javaScript, preferably in some editable form, so that it can be hacked about with PHP or whatever afterwards.

Adobe Edge looks promising, but I am wary of Adobe's ability to manage multimedia authoring tool development. Their track record is abysmal. Can they just not screw up, bloat, and hobble their tools with the limitations of their broader strategy for reaching 'internet marketers'? I am sceptical.

And I am training my JS/HTML5 muscles for multimedia teaching and multimedia content production, because the quality difference between 'the men and the boys' is going to be pretty stark.

~mico and Brennan Young are right

~mico is quite right. Brennan Young's comments are also good.

Different browsers have always interpreted the HTML standards in their own way. There is no reason to celebrate and think it will be different in HTML5.

Nothing to thank Steve Jobs for.... Adobe take decisions independently of him. He voiced his opinion - which does not mean he is resposible as the person for (quoting from above) "THANK YOU, Steve Jobs for giving Flash the boot.". To think that is to think that somehow Jobs had a mind control device actively controlling Adobe leaders to nod to him - something I think people who seem to have found a 'god' in Jobs tend to do with anything Jobs ever said. Or at least that somehow the media turned to all things Jobsian to assert him as right and send Flash down the pan - that isn't right either.

Flash was coming to a natural point of adapt-or-die within its lifecycle. It is now quite long in the tooth. Adobe are acting on this.

As somebody has previously said in another Reg article on HTML5 and Flash: HTML5 is an extension to HTML that brings some new features, but a proper programming language it is not. It is not object-oriented or capable of building applications with the same OOP logic as you could with Flash. OOP exists for robustness, code re-use and instancing. HTML5 does not offer these things. There is a wealth of things that are better accomplished in the old stalwart than would be the case in HTML5.

Anybody who thinks HTML5 or Steve Jobs has killed Flash off is misleading themselves. Adobe appear to be adapting Flash to encompass HTML5 where warranted in order to produce something new (by name and brand) but is actually just a natural product evolution. We will see how Flash-like the result is. I suspect that this 'Flash replacement' from Adobe will see nearly as much use and HTML5 will remain a HyperText scripting tool and never develop to a full language.

I am not an advocate of Flash. I personally think it has its problems. Likewise HTML5 is not a 'magic bullet' for web development.

And if you don't believe me... read this article for starters:

http://www.uza.lt/2010/04/29/my-take-html5-vs-flash/