Microsoft ruffled a few feathers in the browser community this week by seemingly inventing a new term; native HTML. "The sites that you visit and the sites that you write are better when your browser runs them natively," as IE leader Dean Hachamovitch put it.
Some took it as marketing fluff explaining why IE9 doesn't run on XP, others that Microsoft was wanting developers to tie Web sites to Windows and break the open Web. Yes and no. 'Native HTML' means a few different things. It doesn't mean the Web should be tied to Windows; although Microsoft is emphasising the app-like Windows feel sites can get with pinned icons and jump lists, as well as mentioned "native experiences, fitting into the consumer desktop and having jump lists," Hachamovitch also repeatedly emphasised performance; "sites that feel and run more like native applications than Web pages," "the performance and feel of sites, it's just closer to what people expect from apps," and - demonstrating IE10 using 10% less CPU to render a page with ten embedded HTML5 videos in ten canvases faster than Chrome with five - "browsers that run natively with full hardware acceleration deliver website experiences that feel and run more like applications than pages."
Native HTML doesn’t mean the Web is better on Windows; it means the Web is better on a browser that's built closer to the metal - and Microsoft is claiming that IE9 on Windows 7 is the only browser that makes a good job of that (until IE10 comes along, of course). And given that we've seen IE10 running on Windows on ARM almost indistinguishably from Windows on x86 (on a limited and controlled test, of course), Microsoft is staking a claim for native HTML on whatever the next version of Windows turns out to be.
Hachamovitch justified throwing XP under the bus (over and above the out of support point, which explains why IE10 will only run on Windows 7 and 8), also in the name of performance; saying it wasn't worth making a native IE9 for XP and it wasn't worth doing anything less either. "Using compatibility layers to run cross-platform in non-native ways makes browser development easier; it doesn't always make a better browser or the best experience of the Web," he said. "Browsers that optimize for modern operating systems more directly deliver better experiences. Someone who's spread too thin across too many different operating systems and too many different versions of too many different operating systems can't deliver the best experience on any one of them. Building a new browser for the 10-year-old version of Windows that came with IE6 didn't make sense. There are too many limitations in the graphics and security architecture. "
Native HTML is also a putdown about Flash (and, to a lesser extent, Silverlight)."As developers, we'd rather have native support for important features rather than an add-in or a hack," Hachamovitch said. "A year ago, many of these experiences would have required a plug-in or not been possible in a browser… Native HTML5 support in Windows with IE9 makes a huge difference in what sites can do."
That's less of a putdown of Silverlight than of Flash because of the ways Microsoft is using Silverlight outside the browser - on Windows Phone and almost certainly on the forthcoming tablet version of Windows. And that's where 'native HTML' gets very interesting, and might even explain the seemingly throwaway line "Native experiences continue to be the best experiences. On phones, for example, people consistently choose native applications over websites." Assuming that's not just a metaphor like Hachamovitch's nod to the classics (appreciated by us Brits in the audience; "Dante is best in Italian. Shakespeare is best in British. Surak is best in Vulcan.") it could reinforce some Windows tablet rumours - and it's inspired us to some speculation about 'the next generation of Windows' on ARM, as Steven Sinofsky called it at CES.
That may have just been Sinofsky's typical caution, although he certainly has a very successful poker face - having looked us in the eyes at PDC 2009 and professed to know nothing (nothing, I say), about any project to port the Windows kernel to ARM (and we certainly don't fault him for not wanting to reveal a secret project). But he was very definite: "I didn’t say Windows 8, I said the next generation of Windows," he repeated when someone used the version number. And he told us later not to expect exactly the same as desktop Windows, no matter how much work partners and developers did. "We have to do some surveys; figure out how to package it, how to position it. We have to not oversell it, because there will be some capability issues."
The Windows Phone Mango update is the beneficiary of Windows moving to ARM, getting IE9 this autumn because of the work the IE team is doing to put IE10 on ARM Windows. IE10 supports two page layout standards that the IE team is calling mature but hadn't managed to support for IE9 (Flexible box and multicolumn CSS3) along with a grid page layout Css3 proposal Microsoft came up with itself; ideal for laying out an interface that reflows when you rotate the tablet you're viewing it on. And while Vista doesn't come out of support until May 2012, putting the likely release of IE10 for x86 Windows soon after, that's not enough to rule out an IE10 for ARM Windows earlier - and ARM Windows and ARM IE10 will be written as close to the hardware as x86 Windows. They have to be, to get the performance to run Office and get the performance on video and hardware acceleration that Microsoft wants.
If you don't believe that Microsoft can leave the tablet market to iPad and Honeycomb until 2012, native Web with app-like sites could give Microsoft a shortcut. While we still believe x86 Windows 8 will be announced in September at PDC, in beta by Spring and shipping on PCs in autumn 2012, we can imagine Windows Tablet launching at PDC, with no public beta because no-one has the hardware to run it and shipping only on new ARM tablets soon after (something big for CES?); coming with oh-so-native IE10, Office, Windows Live - and with Silverlight apps as the way third-party developers can deliver software quickly. There are 13,000 Silverlight and XNA apps for Windows Phone already and it only launched in November.
What about full Windows apps? Maybe they'd run (Microsoft could court selected developers at the Worldwide Partner Conference this summer and announce developer tools as PDC). Or maybe that would have to wait for an update, coming perhaps at the same time as full Windows 8. Of course a major future feature like that could completely Osborne the launch (Osborne: to kill your product at birth by also announcing the far more desirable successor, as the venerable portable PC maker did), unless Microsoft promised full updates. That would be a lot easier if the Windows Tablet was a badged Microsoft product, although that strategy has only worked for Xbox, where Microsoft has no OEMs to estrange. Although Windows tablets will doubtless come with 3G modules, the operators wouldn't be able to block an OS update coming through Windows Update as a service pack the way they have delayed the first two Windows Phone updates. And while neither Windows Update nor Windows InTune can deliver full OS updates, if you can deliver as significant an OS upgrade as XP SP1 through Windows Update that's not really a barrier. Plus, Microsoft has Azure; it could build a cloud OS deployment service by next autumn, or offer a download to PC, install via USB update. Where there's a will, there's a way.
The two big questions are: is the ARM version of Windows really that far along - and does Microsoft have the will to push out a tablet version of Windows before Windows 8? The reported lack of enthusiasm for Honeycomb tablets might make it less urgent, but with touchscreen support going into Chrome OS and the iPad juggernaut showing few signs of slowing (and RIM promising the missing email and BlackBerry apps for PlayBook soon), Microsoft may not be able to just keep pointing out that Windows tablet PCs do more than their sleeker, sexier counterparts. The impressive NUI future Microsoft keeps hinting at for Windows 8 may be a lot more exciting than tablets when it arrives (every new Kinect demo we see expands the possibilities of interaction in a way touchscreen can't match) but with a 3% drop in worldwide PC sales in Q1 this year 2012 is a long way away.
Mary Branscombe and Simon Bisson
P.S If we seem to be refining too much on a word, that may be true - even a word that stirred the Mozilla community to file a humorous bug about the lack of 'native HTML' support in WebKit and had some of the Chrome team joking about the need to add more Chrome sessions to Google IO to keep up with Microsoft coverage. But it's notable quite how many time that word can up in the keynote. Here are all the places we spotted it…
Native experiences continue to be the best experiences. On phones, for example, people consistently choose native applications over websites. As developers, we'd rather have native support for important features rather than an add-in or a hack. Now, lots of things lose in translation. Dante is best in Italian. Shakespeare is best in British. Surak is best in Vulcan. It's just -- some things are just not the same in translation. And the same is true for the Web. The sites that you visit and the sites that you write are better when your browser runs them natively. Every library, every layer, every abstraction between your site and the device challenge performance, reliability, and the overall experience. Native experiences are the best experiences. Web experiences are the most important experiences. The only native experience of the Web of HTML5 today is on Windows 7 with IE9. To deliver the most native HTML5 experience, we built IE9 from the ground up for HTML5 and for Windows...we'll show examples from the development community of sites taking advantage of native HTML5 support. A year ago, many of these experiences would have required a plug-in or not been possible in a browser… Native HTML5 support in Windows with IE9 makes a huge difference in what sites can do...hardware acceleration is not just a yes or no checkbox. Much like standard support is not just a yes or no checkbox. And the same with native experiences, fitting into the consumerdesktop and having jump lists. Now, many people want a better Web, a faster, more powerful and more native, a more interoperable Web. We want actual progress to that goal, not just activity. The Web makes progress when developers can take advantage of new technology to build sites that feel and run more like native applications than Web pages across production-quality browsers that use the same markup consistently. That's when the Web delivers on the promise and the value of standards — when we as an industry deliver on consumer and business HTML5...Native implementations are just better for developers, for consumers, and businesses. The performance and feel of sites, it's just closer to what people expect from apps. Using compatibility layers to run cross-platform in non-native ways makes browser development easier; it doesn't always make a better browser or the best experience of the Web. The other thing to point out here is that IE running 10 of these videos across 10 different canvases is actually using about 10 percent less CPU than the other browsers. And, again, that's just another example of HTML5 running natively... browsers that run natively with full hardware acceleration deliver website experiences that feel and run more like applications than pages...IE9 delivers native support for HTML5 on Windows. You and your site can take advantage of that today and deliver significantly better browser experiences….we took an early look at IE10 and how it makes progress with native HTML5 support for more Web standard technologies.