The victory of HTML5 as the dominant mobile platform became inevitable last year or even earlier, but still looked as though it would be a gradual process. However, its progress has accelerated, despite lingering battles (such as the Google-Apple dispute over codecs), and now even Adobe has waved the white flag. The Flash company has been investing in HTML5 and preparing itself to coexist with the newer system all year, but its new restructuring ditches mobile Flash altogether after the forthcoming new release. That changes the battle lines once more when it comes to the all-important area of cross-platform development. A powerful, if ageing, contender is removed, confined to the desktop henceforward. Java, Silverlight and others still hang on, but of course all the attention is on HTML5 ‘ and that in itself intensifies the fight among the web players to grab the steering wheel and shape exactly how that standard underpins the future mobile cloud experience, and the competitive dynamics of its providers.
Adobe’s complete conversion to Flash is not surprising in itself, but the rapid timing would have been unthinkable until only recently. Even when it started to build its HTML5 arsenal, with moves like the acquisition of Nitobi, Adobe was still talking in terms of many years of life for mobile Flash, despite the high profile hostility of Apple. However, the shift to the new platform has been ongoing for months, for instance with the summer’s unveiling of Adobe Edge, which incorporated HTML 5 to let developers create simple animations for viewing oniOS and Android.
Adobe says it will stop providing updates to its Flash software for mobile devices and will use HTML5 for its mobile activities after the current release. The company said in a statement: ‘Our future work with Flash on mobile devices will be focused on enabling Flash developers to package native apps with Adobe AIR for all the major app stores. We will no longer adapt Flash Player for mobile devices to new browser, OS version or device configurations. Some of our sourcecode licensees may opt to continue working on and releasing their own implementations.’
As Adobe put it: “HTML5 is now universally supported on major mobile devices, in some cases exclusively. This makes HTML5 the best solution for creating and deploying content in the browser across mobile platforms.” The new browser standard has clear advantages over Flash, allowing web applications to look and behave almost like native apps, and without plug-ins, though it does not yet approach the functionality of full-blown Flash.
But it has many huge companies and millions of developers working to improve that situation, and Adobe could not compete with that level of resource. While Flash was a huge force on PCs, it never achieved the same dominance in the mobile world, where its Lite version was excluded from the iPhone and accused of being a power hog, while the promise to bring the fully fledged version to handsets was beset by many delays. Thus the company has been forced to reverse its previous position that it alone could bring the ‘PC web experience’ to small screens, and most crucially, has abandoned the cross-platform ambitions of Flash forever. Instead, it will have to play out its cross-OS strategies via its tools, adapting them for HTML5.
After Adobe releases Flash Player 11.1 for Android and BlackBerry PlayBook in the near future, it will “no longer continue to develop Flash Player in the browser to work with new mobile device configurations’, though it will keep issuing bug fixes and security updates for existing users. The change of direction does not affect PCs, and the firm is already working on Flash Player 12 to increase desktop functionality as well as enabling “a smooth transition to HTML 5″.
On PCs and Macs, Flash still has the scale to justify the continuing commitment, but this was never true in the mobile world once there was a credible alternative for delivering video and rich web experiences. The fragmented state of the smartphone market requires huge amounts of time and money to be spent on maintaining a variant for each operating system and chipset, especially as these change so frequently. This is a major challenge for any company bringing software to mobile platforms, and of course one of the greatest advantages of HTML5 is that it supports, theoretically at least, ‘write once deploy anywhere’.
Of course, there are downsides to this promise, and not just for Adobe ‘ Apple, though a big supporter of HTML5, also has to face the erosion of its tight control over its applications base and user experience, as developers start to deploy web apps accessed via the browser, instead of native downloads optimized for a specific OS. HTML5 will be a big adjustment for all the players, but at least Adobe is not clinging to its old norms after they are already dead, a mistake it has made before.
One of Adobe’s missteps with Flash was to take too long to bring full functionality to mobile devices. Instead it gave Apple its weapon by relying on the stripped-down Flash Lite, which ‘ as the iPhone maker always claims ‘ became increasingly inadequate as smartphones became capable of PC-class video. Full Flash Mobile did appear, but only after industry attention was drifting to HTML5 ‘ as shown by the failure of Adobe’s own app store, which was meant to make Flash the main instrument to span the conflicting mobile operating systems. By contrast, the new wave of HTML5-friendly Adobe tools ‘ notably the cloud-based Touch Apps, released in October – has been designed primarily with mobile devices in mind, and with equivalent functionality across desktop and handset from day one.
Having set out an HTML5 coexistence path in the summer, the company shifted more clearly towards migration at its MAX event in the fall. The three key elements in that cloud/mobile/web push were Touch Apps, new features for the Edge toolset, and the acquisition of Nitobi, maker of the PhoneGap open source framework, which could smooth Adobe’s transition to its new environment. All of these presaged this week’s news, in particular the Nitobi deal. The smaller organization’s PhoneGap framework has been the basis of web developer kits from several mobile players including Sony Ericsson and Palm and it was a close partner in Symbian. Last year, Nitobi announced PhoneGap Build, a cloud-based tool for creating apps in HTML, CSS and JavaScript, uploading them to Build, and getting them back ready to run on various smartphone platforms including Android, iOS, Windows, Symbian, BlackBerry and bada.
Nitobi’s mission has been to increase the number and usability of web-based products for mobile platforms, by making development easier and cheaper, and over time it believes the functions of PhoneGap, and many other cloud-based platforms, will be folded into the browser. The fact that this thinking has landed up inside Adobe shows how rapidly the larger company is evolving to cope with a post-Flash world. But Adobe’s competitive edge will remain in highly functional tools, more suited than open frameworks like PhoneGap to deliver heavy duty apps, such as games, across different browsers and OSs. While PhoneGap will be the heart of mobile specific and low end media activity, at the high end Adobe will still need key products like DreamWeaver, and will have the challenge of running them across PC and mobile environments, and bridging them with PhoneGap.
Analyst Jack Gold of J Gold Associates believes Flash remains superior to HTML5 for mobile content delivery, but agrees that the resources required to preserve it had become too high and the chances of beating HTML5 too niche. ‘Adobe is making a calculated decision that even with superior technology, betting on mobile content delivery for Flash enabled devices is beyond their ability to generate enough profit for the amount of commitment of resources needed,’ he wrote in a research note. ‘It’s like more popular VHS (or BluRay) winning over technically superior Beta all over again’. Adobe may ultimately gain revenue from the popularity of HTML5 development, but we believe this move signals the slow decline of Flash from the overall market.’
Other rich media platforms may cling on a little longer, though there are question marks over both Microsoft Silverlight and Oracle JavaFX. Microsoft has already placed HTML5 at the heart of its cross-platform strategy and reduced the role of Silverlight, from premier cross-platform runtime to a toolset for creating rich media apps, especially for WP7. The company is about to release Silverlight 5 to manufacturing but there are reports that there will never be a Silverlight 6. Even release 5 may support only Internet Explorer, and Microsoft is said to have reduced the team working on the platform, scaling back the once-ambitious plans to incorporate it as a unifying factor across all its media systems, (Windows Phone, Zune, Xbox 360).
Like Flash, Silverlight will have contributed to more open technologies which will have a longer term role in the web, such as the XAML mark-up language, a de facto standard, and its concepts underpin the Metro user interface for Windows 8, which promises to be a major shift in cross-platform experiences. Andrew Brust, a Microsoft regional director, told ZDnet: ‘It’s pretty clear to me that the principles of Silverlight – including the use of XAML as a mark-up language, C# and VB .Net as programming languages, a streamlined .NET Common Language Runtime profile, packaged deployment over HTTP and a sandboxed security environment – are alive and well in the native XAML/.Net approach to developing Metro-style apps on Windows 8. It may not be not Silverlight to the letter, but it’s Silverlight in spirit and natively supported by the OS to boot.”
Another platform which has heard many death sentences pronounced is Sun’s JavaFX, now owned by Oracle. Released in early 2009, the technology was Sun’s biggest effort to make a serious commercial mobile business based on its Java crown jewels, though a long gestation period, and then a degree of sidelining by Oracle, reduced its impact. The system provides a runtime for deploying rich internet applications on mobile devices and at launch was positioned as a competitor for Flash and various Android and Symbian toolsets. At the time, Sun’s then-head of software, Rich Green (later a Nokia CTO), said that, by releasing a full cellphone software stack, the firm was going head-to-head with the mobile majors.
The most interesting aspect was Sun’s attempt to reunify Java, which had suffered a particularly serious split when Google chose to develop its own virtual machine, Dalvik, for Android, and adopting the Apache open source process which had long been a thorn in Sun’s side. Oracle has now taken up the baton on that attempt with its legal attacks on Google, alleging that the search giant committed Java patent and copyright infringement in creating its own fork.
JavaFX has been a less successful threat to Android’s power than the lawyers, but recent indications are that Oracle is not going to forget the system quietly, but could give it one last chance at powering a serious mobile software push. The database company gave the technology, and plans to open source it, far higher billing than expected at a recent developer event and has also been linked with a possible bid for HP’s webOS (see inset). The latter may be fantasy, but it actually makes sense ‘ the two platforms could be combined to create a full stack well adapted to two key trends, the evolution of featurephones (most of which support Java) into full webphones; and the incorporation of cloud connectivity and web capabilities into vast numbers of devices well beyond the handset.
The dominant software to power these two important developments is still undecided, with Google taking the lead with Chrome OS and other tools, but many other cloud-oriented candidates emerging ‘ Mozilla Boot to Gecko, Facebook’s putative browser/OS, Baidu from China with its ‘box computing’, the Intel/Samsung Tizen effort, a possible full platform from Amazon, and whatever may evolve from Microsoft’s embedded and Metro initiatives, and from the labs of Apple.
So can Oracle give Java a starring role in these developments, a role for which it looks well suited by virtue of its massive installed base, developer support and cross-platform nature. Sitting on top of a thin Linux cloud OS (such as webOS) and surrounded by a full web toolset and mobile stack, Java could provide all the elements which Google, Facebook and the others are racing to assemble.
Hence why JavaFX could make a comeback. It provides an integrated platform combining basic OS, browser, user interface and other key elements, all delivered in ‘pure’ Java and underpinned by the FX scripting language. At launch, this held out the promise of building a rich media platform suited to the midrange, low cost handsets where Java still holds the fort as an apps standard (it lives on 1.8bn handsets, increasingly using at least rudimentary app stores and interactive content). However, almost three years later, Android may have got there first on conventional phones, with traditional Java under pressure from Android’s move into the mass market. However, Android has its limitations when it comes to embedded and ultra-low cost devices, which may see a new push by Oracle to make JavaFX relevant.
The key promise of JavaFX when it was mooted in 2007 is now a familiar one, but was more radical then – to remove the need for a fat, resource hungry OS from low end handsets, while still supporting rich media and web apps, which underpin mobile usage growth in emerging markets just as elsewhere. This was Sun’s big chance to turn Java into a strategic mobile platform, at a time when the major smartphone OSs were ignoring the mass market and battling for the premium users. The world has changed now and all the majors know they will need a slimline OS, probably Linux-based, which can be embedded in the tiniest devices and access the cloud constantly and efficiently. Hence Google Chrome OS, Tizen effort, and HP’s tortured webOS. JavaFX can sit on top of Linux as well as native Java alone.
The past week has seen Oracle named as an interested party for webOS, and outlining its plans for JavaFX. Seeking a broader base, it will open source the platform and replace any closed code with open code, the firm promised. This will make JavaFX into a stepping stone for the next generation Java client toolkit, geared to an ever-widening range of web client devices. Richard Bair, chief architect for client Java, said: “Our basic motivation for wanting to open source JFX is to build community and ecosystem support and adoption around JavaFX by increasing transparency.’ This will make the broader open source Java project, OpenJDK, “bigger and richer’, according to analyst Al Hilwa of IDC. “This shows signs of investment from Oracle and ‘ also has the potential to broaden the client Java technology if contributors come on board and improve the code.”
The most symbolic gesture that Oracle aims to give JavaFX a new role in the limelight came last month when it demonstrated the technology running on an iPad, an apparent provocation to Apple, which has not permitted Java to run on its iOS products. This raised the specter of a rerun of the Apple-Adobe stand-off, this time concerning Java ‘ which would certainly give the latter a new prominence in the minds of developers and mobile planners. And if JavaFX is seen gaining a strategic position in the new wave of cloud/browser computing, Apple could have a tougher fight this time, especially if Oracle succeeds in its bid to gain broad acceptance for FX across Windows, Android and other platforms, not just basic Linux. As with Flash, Apple’s argument for excluding JVMs from iOS is that it would end up with a ‘lowest common denominator’ performance, with interpreted languages unable to deliver the level of functionality Apple demands for its user experience. Oracle argues that a rising number of App Store products are getting round Apple’s mandate on its Objective-C language by linking the JVM into the program, as it did with its FX demo. It has also worked on other ways to connect Java and iOS, notably its Project Avatar for dynamic rich clients, which links HTML5 and Java to the Apple system. Avatar enables hybrid apps where HTML5-based UIs share content between Java clients and Java EE servers, in data centers or the cloud.
Of course, making JavaFX work with HTML5, rather than starting a fruitless battle to provide an alternative, will be critical to Oracle, and it will have seen, in Adobe’s failed hybrid Flash/HTML5 attempts, how hard and expensive this will be in the mobile and cloud client worlds. However, Oracle has deeper pockets than Adobe and Java has an even greater entrenched grip on the mobile market than Flash. This year’s JavaOne event showed Oracle becoming far more open to HTML5 and other popular web technologies, for instance by opening JavaFX up to any JVM language rather than insisting on its own JavaFX Script, a major obstacle to programmer adoption. Most importantly, Oracle is no longer ignoring HTML5 and JavaScript. JavaFX 2.0 will allow HTML5 content to be embedded in JavaFX applications and be rendered via WebKit.
And while Oracle may have been criticized by the Java community for sidelining mobile efforts since it acquired Sun, it knows it still has a powerful weapon in terms of its handset installed base and the broad use of the Java language. This has actually been stimulated by Android, despite the falling-out of Google and Oracle, according to Tiobe Software, which runs a monthly study of the most popular developer languages. Java remains at the top of the tree, being used by almost 18% of developers, just ahead of C. ‘Java is definitely growing in the mobile market,’ said Paul Jansen, Tiobe’s executive director: “The Java-based Android platform is pretty popular among developers.” After C comes C++, Microsoft C#, PHP and Apple Objective-C for Mac OSX and iOS, which saw the highest growth in usage, up 2.79 percentage points year-on-year to 5.98%. Still a minor player, but rising quickly, is Google’s new Dart language, which it has launched with the express purpose of addressing weaknesses in JavaScript for web/cloud apps. It has risen from 107 in the league table a month ago to 63. Project leader Lars Bak says the language is suitable for everything from small, unstructured projects to large, complicated efforts such as Google Docs. ‘If we want to focus on making the web better over time, we have to innovate,” he said in an interview, and that means rethinking popular programming languages. However, he pulled back from internal memos leaked last year, which said Google aimed to make Dart (then called Dash) a replacement for JavaScript ‘as the lingua franca of web development’. Bak insisted: “It’s not going to replace JavaScript. JavaScript is a cornerstone of the web today, and it will continue to be for a long, long time.” The next important step for Dart will direct integration into the Chrome browser to improve web apps performance, and a key goal will be to ensure high performance on ‘all modern web browsers and environments ranging from small handheld devices to server-side execution’. There will be various modes geared to different types of device and project ‘ Dart will be able to run without a virtual machine, for instance, and there will be a compiler that translates Dart code into JavaScript code for browsers that don’t support Dart.
This is another example of Google using the Chrome browser to get its technology deep into the heart of the web programming environment with a view to controlling de facto standards. Its WebM video codec ‘ battling with the Apple-backed MPEG LA alternative for the key role in HTML5 ‘ is another instance, initially pushed by being integrated into the increasingly popular Chrome. Building Dart into Chrome could let Google build Dart versions of its advanced web apps which could outperform JavaScript versions, and these could be showcased in its own browser. HTML5 may be almost universally supported as a web apps standard, but that does not mean that powerful players will not fight to have a stronger grip over how it evolves, in terms of patents and developer influence.
Google is pole position here, but Facebook is determined to leverage HTML5 to gain its own place in the web apps sunshine. It has spent this year making acquisitions and announcing products to turn its social networking system into a fully fledged cross-platform offering which can work as a native app, a cloud service and a thin featurephone function; and which will evolve into a fully fledged developer platform centered on HTML5 and the browser/OS. The latest acquisition to support this ambitious strategy is mobile tools start-up Strobe, which was born out of SproutCore, an open source JavaScript framework built by Charles Jolley in 2007 while he was an engineer at Apple. Strobe is an app delivery network which supports SproutCore and helps get mobile HTML5 apps up and running quickly, and into web app stores. According to a Strobe blog post, the company ‘was founded on the belief that HTML5 can transform the way average people use their mobile phones through apps that are available everywhere, anytime, on any device’. Importantly, as well as harnessing increasingly powerful mobile browsers, it also offers readers which enable the coexistence of HTML5 and native apps, providing a common user interface for either across multiple operating systems.
There will be many such attempts coming from the web players, which may lack the installed developer toolkits and APIs of Oracle or Microsoft, but have alternative power bases in the browser or key web services. Another example is Mozilla, one of the open source players which is seeking to replicate PC-based success by creating a mobile platform. Earlier this year, the Foundation talked about its plan to develop one of the new breed of browser operating systems, rather like Google’s Chrome OS, with its ‘Boot to Gecko’ project. Now it has published a timeline and published the first images of its planned user interface, with developers able to test the new system as early as December.
Boot to Gecko (B2G) initially combines the browser UI with a stripped-down Android but is likely to move away from the Google system in future and target even thinner Linux variants (the system platform is called Gonk, and comprises the Linux kernel and some low level userspace libraries derived from Android, but will not include any of Android’s Java libraries or stacks). The project’s eventual goal is to build a framework that will let apps run from the web on any OS which supports the B2G technology. The mobile OS will reach developers before year end, with product demonstrations planned for the first quarter of 2012 and full release around mid-year, according to the project’s web site. Tasks include the development of an ereader, media player and browser, all tailored for the new platform but open to all kinds of web content and services.
Like Chrome OS, B2G is designed for devices which mainly behave as thin clients, constantly accessing the cloud rather than having significant local processing and storage. As in other such initiatives, HTML5 is central. Mozilla plans to build a new HTML5 WebAPI to power most of the functions of the phone, using JavaScript APIs for real time communication capabilities in order to enable apps to render in the browser. The browser maker also plans to create an Open Web App Store and suggest that, by purely using HTML5, users could demo open web apps without any sort of installation or software upgrades.
This is a far ‘purer’ web approach than those of companies living with native apps legacies, like Oracle or Adobe, and that could be its downfall ‘ Google would instinctively like to push harder to move away from native code altogether, but as its Android strategy shows, has acknowledged it needs to be more pragmatic and accept that local, fat OSs and apps will survive for many years to come, especially on mobile devices with unreliable connectivity. However, a key role for the open source community is often to show the way the world could or should go, and set a high ‘ if idealistic ‘ bar. Such visions are already creating major repositioning by the old guard, as seen in the Adobe U-turn. So far, Google still looks like the giant with a foot reasonably firmly balanced between all camps, but it cannot ignore the web-focused upstarts ‘ nor, less sexily, a resurgence by the old dinosaurs of enterprise software, as represented by Oracle/Java. Perhaps we should not forget that Oracle and Sun were the first major vendors to challenge the Microsoft ‘fat PC’ model, back in 1996, with the concept of a slim, always-on ‘network computer’, now reborn as a cloudbook. That vision was stillborn at the time, but it led to 15 years of intensive R&D at Oracle. Riding the wave of ubiquitous connectivity, that investment may finally bear fruit now.
Oracle could bid in webOS fire sale, says report
Hewlett-Packard still hopes to offload webOS and regain at least part of the $1.2bn it paid for Palm, though the process is starting to look like a fire sale. According to a Reuters report, HP is now looking for a price in the “hundreds of millions of dollars’, and we can be sure that will fall in the low hundreds, a very poor return on the Palm buy and the investment put into developing the failed TouchPad tablet. Oracle and Amazon are said to be among the interested parties.
A long list of potential buyers has been linked to webOS since former CEO Leo Apotheker decided to spin or sell off HP’s PC and mobile device activities, but no bids have materialized. They have included , at various times, Facebook, Samsung, HTC, Qualcomm and LG. The current candidates are said to be Oracle, Amazon, IBM, Intel and RIM.
The platform has seen a dizzying flipflop of changing directions this year ‘ first placed at the heart of an ambitious strategy to dominate the cloud, with webOS embedded in a huge array of connected devices; then apparently axed along with the Pre and TouchPad hardware. Since then, different executives have, at different times, hinted that the former Palm platform would be retained for printers or cloud devices; closed down, or sold. New CEO Meg Whitman has reversed the decision to get out of PCs, but has remained silent on webOS, except to say its fate would be decided in the next few months.
She told investors that HP ‘needs to be in the tablet business’ but indicated that Windows 8 would be the primary platform for that. The Reuters report, which cites four unnamed sources, says HP is being advised by Bank of America Merrill Lynch on the transaction, and an HP spokesperson said: “We are exploring ways to optimize the webOS software.”
Of the rumored interested parties, Amazon is the most long standing and the one with the clearest motive to acquire its own operating system, and one focused on the kind of web/cloud services environment which will underpin the retailer’s growth. It has adopted Android for its Kindle Fire tablet, but a version so heavily customized as to be almost unrecognizable, so its own OS, more naturally geared to online behaviour, would be only a small step further.
Of the other firms mentioned in the report, both RIM and Intel have surely had enough of trying to establish their own cloud-oriented operating systems for connected devices. RIM is struggling to come up with a convincing story for how it will build a bridge between the ageing BlackBerry OS and the new QNX, while Intel has stepped back from its MeeGo Linux-based slimline platform and merged it with LiMO to form Tizen. Another OS with virtually no device or developer base would hardly help matters for either firm.
IBM and Oracle would presumably want webOS for strategies more similar to the ubiquitous cloud vision briefly espoused by Apotheker, incorporating it into any kind of end product in order to create a huge developer base, and an embedded device platform optimized for the vendor’s cloud platform and services. Oracle might have the added interest of boosting its patent portfolio, as it pursues Google over alleged infringements of Java IPR, and even of combining webOS with Java technologies to give it a more modern mobile platform.