Distilling market noise into market sense
The VisionMobile blog is a space where VisionMobile analysts and industry insiders exchange views on the fast-changing mobile market and the trends that define the future direction of telecoms.
The Flash vs. HTML5 Endgame
[In the debate of Flash vs HTML5, has the death of Flash been over exaggerated? Guest author Guilhem Ensuque peeks through thick layers of hype and facts to predict what the future holds for the mobile web].
The last year has seen a flurry of announcements and debate around the rise of HTML5 and the fall of Flash. Some have even gone as far as declaring a “war” between the two, and predicting the “death” of Flash as the outcome. However, as Mark Twain once famously said: “The rumor of my death is an exaggeration”. As we’ll see, the jury is still out as far as the fate of Flash and Adobe are concerned.
A brief (abridged) history of the web
“HTML5” is the new high-tech industry darling, and not just in the mobile space. It has become a catch-all phrase with little meaning when taken out of context. Before we dig into the debate, it’s worth looking at what is HTML5 and where has it come from.
“HTML5” when used as a shorthand, covers of family of web technologies currently being standardised by the W3C and at various implementation stages by browser vendors. The “5” comes from the version increment in the W3C spec number: currently most of the content you read on the web conforms to the HTML specification version 4.01.
To understand what has driven the creation of this new version of web standards, we need to look at the evolution of the web in past years.
In the 1990s the World-Wide-Web emerged from academia to become the ubiquitous medium to share digital documents over Internet Protocol networks. The HTML4 spec was matured in that era, and has been very much geared towards read-only, document-oriented description and hyper-linking. HTML4 mixes typographical tags with document structure description, within the bounds of static pages and has limited support for script-driven page logic and forms (does anyone remember CGI?). In that era of the web, support for multimedia content was notably absent from the web specification; leading to heterogeneous plug-ins striving to provide video delivery in web pages (remember Real Networks? or having to choose the speed of your modem?).
In the 2000s, the web evolved towards more interactivity with the advent of the “Web 2.0” (yet another buzzword) and user-generated content, especially videos uploaded and then streamed over faster ADSL connections. However, the HTML spec did not fundamentally change (apart from an attempt by the W3C to migrate to the stricter XHTML syntax which has seen mixed results in terms of adoption). To cope with HTML4‘s inefficiencies in allowing designers and developers to create interactive “experiences” (i.e. not just documents, but bi-directional “applications” living in your web browser) a number of innovations were introduced :
- clear separation of page structure in HTML (through heavy use of <div> tags) as well as typoraphy and style in CSS (through an arcane and verbose syntax), leading to more pleasant user experience and richer page contents
- PHP-scripted and database-powered back-end logic bolted on top web server systems. This e.g. allowed template-driven content management systems like WordPress and Joomla to rise to prominence, fueling the blog revolution.
These innovations brought the ability to present vast amounts of data in pretty-looking dynamic web pages which mash-in RSS feeds, emails, blogs, Facebook updates, and tweets, and bringing web pages a step closer to applications.
In that era, Flash (or rather the Flash Player) rose to become a ubiquitous browser plug-in for animated graphics and video. At the same time, Flash evolved to provide an out-of-browser Rich Internet Application platform with the AIR runtime and the Flex framework, albeit at a much lower penetration level than the in-browser Flash Player.
We are now at the dawn of the 2010s, and the overhaul of the HTML4 spec is long overdue. HTML5 aims to bring back into the core spec of the web the “side” developments of the previous era and improve on them with a heavy focus on web applications. It also aims to lay the foundations enabling the delivery of web content through a new medium: mobile devices, and ultimately the “Internet of Things”. That history is yet to be written, but we can now ponder about its beginnings and the future.
So, What is HTML5 Really ?
In the context of this new era, the “HTML5” shorthand refers to a family of web standards and browser technologies that span a range of topics:
- A richer styling language: the Cascaded Style Sheets v3 specifications. CSS3 is now famous for its ability to create rounded corners, but more importantly includes so-called “transforms” allowing graphical effects like moves, rotations, gradients, etc. as well as 3D graphical objects manipulations. Much effort as been put by browser vendor to support hardware acceleration for CSS3 rendering. However, the standard is not yet mature and today requires using prefixes specific to each browser.
- Miscellaneous additions catering for the Semantic Web (microdata), security (cross-domain HTTP requests), and more.
Why the clash with Flash ?
There’s no denying that the capabilities brought forward by the emergence of the HTML5 “family” bring browser runtimes on a par with core capabilities of the Flash Player, which if adopted widely could make Flash redundant.
In the eyes of most mobile industry observers, the delays in bringing out a fully-featured Flash Player with acceptable performance on smartphones have played in favour of HTML5. Remember that, as of today, Flash Player v10.1 is only available for high-end smartphones that run the Android version 2.2 operating system. I would estimate that these represent only 1% of the overall smartphone shipments in Q2. This is a far shot from Adobe’s self proclaimed goal of having Flash shipping on 50% of smartphones by 2012 (see my previous article on this topic).
Company Browser / OS HTML5 compliance Nokia Symbian S60 5th Ed. 7% RIM Blackberry v5 0% RIM Blackberry v6 (Torch)* 69% Android v2.1* 50% Android v2.2* 59% Apple Safari for iPhone (iOS 4.0)* 62% Microsoft IE Mobile (Winmob 6.5) 0% Opera Opera Mini (on iPhone) 9%
Figure: HTML5 compliance of mobile browsers
[some notes on the methodology: HTML5 compliance was carried out using html5test.com. (*) denotes a WebKit-based browser. The Nokia Symbian S60 browser, albeit based on an old version of WebKit, scores poorly in HTML5 compliance tests. I could not test Mozilla Fennec, Palm’s WebOS browser, nor Opera Mobile.Opera Mini is a special case due to server-side rendering.]
Making things worse, Apple has stayed firm on its policy to not allow the Flash Player browser plugin on its iOS devices (iPhone, iPad and iPod Touch), preferring to rely on its in-house video streaming capabilities developed within its HTML5-capable WebKit browser core and QuickTime player. And to make things even more complicated, Steve Jobs’ “Thoughts on Flash” have played a key role in fanning the flames of the “Flash is dead, long live HTML5” fire.
Moreover, Google’s Android, Palm’s WebOS and, more recently, RIM’s Blackberry also embed web browsers based on WebKit that score very high in terms of HTML5 compliance, as can be seen in the table above.
Thanks to WebKit, half of the smartphones being shipped are poised to have the Flash-like capabilities brought by “HTML5” built into their browsers. However, let’s not rush in declaring Flash “dead” and Adobe a company in decline as a result.
Does HTML5 matter to Adobe ?
HTML5 is actually good for Adobe’s business. Indeed most of Adobe’s revenues do not come from Flash as can be seen by breaking down the Flash product portfolio::
- The Flash Player: the plug-in is free and is therefore represents an R&D cost for Adobe. No impact there. One might argue that, if HTML5 were to totally eliminate the need for the Flash Player, it would the positively impact Adobe’s bottom line in the unlikely event the company were to lay off the entire Flash Player team
- The Flash “Platform”: “auxiliary” products that rely on the Flash Player include the Flash Media Server and Flash Access product ranges, licensed to organisations that use Flash to deliver streamed video content (e.g. Hulu, Influxis, Brightcove). The “Platform” also includes the commercial Flash Builder IDE allowing the development of Rich Internet Applications (and the associated free and open-source Flex framework). As can be seen in the chart, these represent a minute proportion of Adobe’s revenue. As we will see further down, these products are not going to disappear overnight due to the emergence of HTML5.
However, HTML5 does put competitive pressure on the product management and engineering teams responsible for the Flash Player to out-innovate the evolutions in browser technology. Adobe points out that this is “business as usual for them” as –they say- it was never their intention to fully replace the browser altogether, but rather complement its capabilities with innovative features, and harmonise areas in which standards have been implemented in an inconsistent fashion across browser runtimes.
As an engineering-driven company, Adobe aims for Flash to stay one step ahead of HTML5 technology implementations, as it already is today in numerous areas. Indeed, an agile R&D division within a single corporate entity will always be faster than a “snail driven by a committee” as the W3C HTML5 spec bodies have been dubbed by some.
Some areas where Adobe is pushing the envelope for the Flash Player include 3D rendering with hardware acceleration, concurrency support, IP TVs and peer-to-peer media delivery. The latter is an interesting transposition of the file-sharing P2P concept; imagine tens of millions of users watching the same live video coverage of the opening ceremony of the 2012 Olympics in London. No server farm or CDN today is capable of sustaining such a peak demand. By allowing instances of the Flash Player across millions of peers to share chunks of the video stream at the edge of the network could be the answer to the problem.
Beyond innovation, another aspect to factor in is that HTML5 is still in its early stages of implementation across browsers, with Microsoft’s uber-popular Internet Explorer browser today lacking any form of HTML5 support whilst representing close to 60% of the web user base (see chart below). Even with the IE9 beta improving HTML5 support and other browsers consistently gaining market share it will still take some years before HTML5-capable desktop browsers dominate the installed base. This will justify the existence of Flash in the desktop browser space for years to come and give some leeway to Adobe’s engineering teams in designing more innovative capabilities.
Company Browser HTML5 compliance Microsoft Internet Explorer 9 beta 32% Microsoft Internet Explorer 8 9% Microsoft Internet Explorer 7 4% Microsoft Internet Explorer 6 0% Mozilla Firefox 4 beta 5 68% Mozilla Firefox 3.6 46% Mozilla Firefox 3.5 42% Chrome v6 72% Apple Safari v5 69% Opera Opera browser v10 53%
Reality check: comparing Flash and HTML5 in key areas
So how is Flash vs HTML5 faring today? For review purposes we can single-out a few key areas of Flash and HTML5 competition, specifically display advertising, video delivery, games and application development.
Display Advertising: a slight advantage for Flash
One of the main use cases for Flash (and big source of annoyance to web users) is display advertising. “Display” adverts are animated banners that appear at the top, side or overlaid in front of the web content you. As annoying as they may be, display ads are a necessary evil for the online world since they represent 40% of the revenues that the digital content and e-commerce ecosystems live on. Even Google uses Flash in its DoubleClick Studio rich advert SDK for advertisers.
Some have said that because HTML5 will kill Flash, those annoying ads will disappear. I would rather think that they may be replaced by equivalents designed in HTML5/CSS3, with the caveat that they may look crappier in most of today’s browsers than their Flash counterparts, as can be seen from these examples.
Indeed a point often overlooked is that today’s HTML5 graphical rendering capabilities are at the level of what Flash capabilities were some years ago and CSS3 transforms allowing to design good “eye-candy” are inconsistently supported across browsers. Therefore I would argue that advertisers will hold back from using “HTML5” for display ad creation in the medium term. The lack of proper HTML5/CSS design tools will also delay this technology adoption by design agencies and creative professionals especially within industry circles where Flash is deeply entrenched.
On mobile devices, the situation will be no different. The blue legos now seen on iPad and iPhones may soon be replaced by HTML5 counterparts; or even by iAds. However, as of today, Apple is the only company creating iAds (in the process levying a hefty ad tax) and is reported to be struggling with the demands of advertisers with its in-house HTML5-based ad creation tools and technologies.
Video Delivery: advantage for Flash
Another area in which “HTML5” has been touted a “Flash killer” is online video delivery. Let’s have a look. As far as basic video playback is concerned, Flash and HTML5’s <video> tag provide the same capabilities, so why not ditch Flash and avoid to end users the (relatively minimal) hassle of installing a plugin?
The situation is not as simple as it sounds as the various browser vendors do not yet all support the same video codecs. On one side, Apple and Microsoft are proponents of H.264; Google is pushing its opensource WebM codec (formerly the proprietary VP8 codec that it inherited through the acquisition of On2/Sorenson); and Mozilla and Opera by default supporting the free and opensource Ogg Theora.
This poses a challenge to online video publishers like YouTube since they then have to re-encode their content multiple times to support each codec.
To end users, this means that videos may not be available in the format supported by their browser. Flash on the other hand, even though it requires videos to be packaged in the FLV container format (not to be confused with encodings like H.264), is available across all desktop browsers and is used as a reliable fallback by “HTML5” web developers i.e. for the 50% or so of IE end-users whose browser can’t render the <video> tag.
Furthermore, the Flash Player supports advanced capabilities required by online publishers such as DRM protection (crucial for pay-per-view business models) and picture-in-picture overlay of multiple video sources with alpha-blending (e.g. for e-learning or overlay of contextual adverts). These capabilities may not be offered for years with the <video> tag in HTML5 browsers.
Casual Games and Visualizations
Flash is the technology that powers some massively popular “casual games” (such as Zynga’s Farmville or Mafia Wars) played by millions of Facebook users worldwide. It also powers numerous other Facebook applications. There was earlier this year a rumor that Zynga was converting its titles to HTML5 to be able to run on the iPhone and iPad. This turned out not to be true, as it announced at Apple’s WWDC that it had ported Farmville to the iPhone as a native app; which may be interpreted as a sign that “HTML5” was not up to the task.
Another area in which today Flash is massively popular is that of visualizations and generative art. There is a large and enthusiastic community that has turned Flash animation into a true art form. Artists like Erik Natzke or Yugo Nakamura (of the Tha agency) are prominent examples of this community. To date, I have not seen any such artistic usage of “HTML5” technologies.
Applications Development: a draw
Web app development is another technology domain where the HTML5 family of technologies has been contending with Flash.
More recently, these frameworks have been forked into mobile variants: JQuery Mobile, Sproutcore Touch and Sencha Touch. Sendra is actually a case in point: the developer company raised $14 million in venture capital, a testament to the significant size of the business opportunity, and has jokingly proclaimed “The End Of Native” (see photo).
This diversity further plays in favour of Adobe’s own web applications platform AIR (a sibling to the Flash Player) and the associated Flex framework, which uses the Actionscript programming language and allows XML-driven UI design through its MXML language.
In conclusion, the momentum behind web applications thanks to “HTML5”’s core capabilities and associated frameworks may seem unstopable, especially as it is driven by technology behemoths like Google and a large enthusiastic community. However this optimism is mitigated by the lack of developer productivity and the rising popularity of Adobe’s application development technologies.
What of the Future ?
Based on the earlier analysis, Flash is far from dead today. There are many cases in which Flash will continue to offer a better alternative (worst case a very useful fallback) to “HTML5” technologies due to the fragmentation in new web standards browser support.
To the question : “will HTML5 kill Flash?” there is no single answer. It all depends on which use case is considered and in what timescale.
On the desktop front, it is the lack of HTML5 capabilities in IE8/9 and their immaturity in all other browsers, that will secure the future of Flash in the medium term. At the same time, Adobe is under pressure from Microsoft, Google and Apple who are betting huge R&D budgets in the development of HTML5-capable browsers and who should be able to out-innovate Adobe in the longer term.
Where would you place *your* bet?
[Guilhem Ensuque is Director of Product Marketing at OpenPlug. He has more than twelve years of experience in the areas of mobile software and mobile telecoms. Guilhem was a speaker at last year’s Adobe MAX conference. His favorite pastimes (beyond mobile software strategy!) include making his baby daughter smile and sailing his Hobie Cat with his girlfriend. You should follow Guilhem on twitter @gensuque_op]