As one peers more closely at HTML5 standards and developments, three conclusions come to the forefront:
Desktop primarily/only apps will continue to coexist with the Dual Web+offline systems; but these desktop apps will be concentrated into math modeling, simulation , and creative design areas that need enormous memory, storage and/or processing speed. HTML5’s message, websocket and geolocation libraries will bring social-media-like facilities to all aspects of business, education, health and social welfare. Microdata, RDF, and the new named <div>s [think <header>, <footer>, <figure>, <aside>, etc] will make scanning, searching, recapturing/repurposing and therefore re-integrating Web data and pages into an organization’s total information management more efficient and/or practicable. Finally, HTML5’s workers will provide background links to the Cloud and/or enormous desktop computing resources ensuring again that those computing resources can be drawn upon from any Web enabled and permitted device. Clearly HTML5 , if allowed to mature and standardize openly, will change IT and computing profoundly.
2)HTML5 standards and also the browser implementations are far from complete – yes there are large islands of reasonably settled standards. But other HTML5 standard sections like multi-touch gestures, the audio and video codecs, Storage and Web Databases, plus microdata and RDF among others are far from settled. Now onto this confusion add the varying browser implementations of HTML5 [that is what this posting is all about – see below] and developing using HTML5 may be “magical … wonderful … and the emerging direction” to Apple’s CEO; but also dangerous and full of bleeding edge hazards to ordinary Web developers. “Danger, Danger Web Will Robinson!”
So clearly it bears repeating – HTML5 is a very big deal; but also HTML5 is far from ready for prime time. So without further to-do, here is the current status of HTML5 Implementations.
Implementations of HTML5 – Summer 2010
The following references and links will try to cite data about the five major browsers [Chrome, Firefox, IE, Opera and Safari]and their HTML5 implementations. This is complicated by the fact that IE6 through IE8 support a tiny fraction of HTML5. So Google has gone out and again done the work for Microsoft with its Google IExplorer Enhancer which increases IE’s HTML5 support markedly. And Redmond has begun serious work itself on HTML5 in its new IE9 version which is available as limited demo here. Unfortunately, many of the web testers have mixed IE with Google Enhancer, IE9 and IE8/IE7/IE6 without Google Enhancer – so the IE results really await a solid beta or later release of IE9.
The second thing Keep an Open Eye has been looking for is an overview of the HTML5 standard implementation – and the HTML5demos.com site has supplied that with a good graphic overview of test pages for many many key sections of HTML5:
HTML5Demos provides not only useful demos of HTML5 techniques using a broad set of the APIs but also tips on the coding. Another website, findmebyip.com, provides similar overview of HTML5 but also more details on CSS3 support as well. Again a very good graphic overview for all the browsers tested is provided:
Also there is an individual browser rating test at html5Test.com. This does not use the popular Modernizr test script from MIT but rather uses its own set of tests and bonuses somewhat like the Acid3 test for HTML4, DOM and CSS. This is a convenient way to see how the particular browser you are using measures up to HTML5:
Finally here is another website testing HTML5 support, When Can I Use, that allows the user to customize extensively what information will be reported. Notable is the futures – what the browser vendors plan for the the end of 2010 and longer term for 2011. This provides a useful perspective on the rate of browser support and adoption of HTML5.
But there is a downside to all of these tests. Like the Acid 2 and 3 tests, they only supply a snapshot of HTML5 compliance – and simply are not comprehensive. Unfortunately W3C has a)a standard that is still evolving and b)a set of small but not comprehensive tests that are a little clumsy to get at and a year out of date . This has some serious consequences. One of the most obvious is claims by vendors like Apple with their Safari 5and Microsoft with IE9 that they are closest to delivering HTML5 to end users. Only two words are required – Caveat Emptor.
What emerges here , despite the lack comprehensive testing of HTML5 web pages, is that the implementation of HTML5 Standards is still quite mixed among the browsers . Yes, there are lagoons of relative stability where developers will find solid browser support. But the browser usage trends still militate against trying HTML5 on the open Web. IE still has 51% market share – so roughly half your users will be unable to use but the most elementary HTML5 without help from Google’s Enhancer on all your delivered Web pages. But then you will have to do careful testing to ensure that a)Google Enhancer works for the HTML5 you are trying to deliver and b)that it does not interfere with other browsers HTML5 support.
Even with the delivery of IE9 in the Fall, developers will not necessarily have big help from Redmond. IE8 still has less than half of IE total usage after 1 1/2 years on the market. And so IE9 will not swamp the market because it only works in Vista and Windows 7 leaving Windows XP users, over 60% of all Windows users, with no access to an HTML5 browser from Microsoft. And unfortunately, corporate IT shops have been reluctant to switch out of IE [IE is tied pretty tightly to Windows system admin convenience]. So unfortunately, HTML5 will be largely at the bleeding edge for the next 1/2 year or so. Thus, Keep an Open Eye promises to do an update to this post in early January 2011.