It's been a work in progress for years, but there are a few more years to go yet before the next version of Hypertext Markup Language is finalized.
Specifically, the World Wide Web Consortium's HTML Working Group is set to announce today that it expects to anoint HTML5 as an officially recommended standard in the second quarter of 2014. That drawn-out schedule contrasts with another effort to make HTML a more fluidly updated "living standard."
"We started working [on HTML5] in 2007," Philippe Le Hegaret, the HTML activity leader for the W3C, told CNET. "We're targeting seven years for completing HTML5."
HTML5 will become the first new revision since HTML 4.01 was released in 1999. Among the features in the next-generation Web page description language: built-in video and audio, a "canvas" element for two-dimensional graphics, new structural labels such as "article" to smooth programming, and a codified process to consistently interpret the hodgepodge styles of real-world Web pages, even when improperly coded.
That doesn't mean interested parties won't be able to employ the new technology until 2014, though. On the contrary, key phases of the coming years' development involve getting feedback from real-world use that's already well under way and ironing out wrinkles that may arise implementing the standard in Web browsers.
At the same time, work continues on a broad range of HTML standards--geolocation, offline data storage, background processing, a direct browser-server communication conduit, and more--that aren't strictly speaking part of HTML5. And after the W3C releases the first "last call" draft of the standard in May--the point at which the W3C thinks the standard's features are set--the W3C plans to begin tackling the early stages of what it's calling HTML.next for now.
Clearly, then, the W3C isn't idling while browser makers and Web developers aggressively push ahead. But the W3C's schedule contrasts sharply with the speed at which the Web is developing today, growing beyond its role as a medium for static documents into a foundation for sophisticated applications. But the schedule also is not a great surprise given the complexity of HTML, the technological and political wrangling among the 55 organizations in the group, and an interest in HTML that's broadening beyond browser makers and Web programmers.
"When you want interoperability at a global scale across a broader industry, it takes time [and] more investment than single-platform stability," said Ian Jacobs, head of W3C marketing. For example, although the Web began as a phenomenon on personal computers, it's becoming a reality on mobile devices and another domain, TV, is coming, as exemplified by a recent W3C workshop dedicated to the subject.
"The key thing here is that there are lots of stakeholders, some of whom may not move at the same speed. One of the pieces of feedback from the TV and Web workshop is that TV manufacturers expect a shelf life of 7 years," Jacobs said. "Because the W3C has as its mission to make the Web available to everybody, we always have to take into account the multiple needs of multiple audiences."
The WHATWG's living document
Even as the W3C proceeds methodically, though, another group involved in developing HTML is changing its philosophy to an even more fluid arrangement. The WHATWG--Web Hypertext Application Technology Working Group--began work on what became HTML5 in 2004 when the W3C declared that the 1999 update to HTML4 was the final version and that the future lay with an incompatible standard called XHTML 2.0. That proved to be largely a dead end, however, and the W3C resumed HTML work in 2007 and now has phased out work on XHTML 2.0.
The WHATWG got its start as an open mailing list, but its founders and decision-makers all came from browser makers--Opera and Mozilla to start, with Apple joining later. HTML governance now essentially involves both the W3C and the WHATWG. One key figure is Ian Hickson, a former Opera and now Google employee who serves as an editor of the somewhat divergent versions of HTML maintained at both the W3C and the WHATWG.
In January, Hickson declared that at the WHATWG, HTML has now become a "living document," a specification that is constantly updated according to need. Abandoning version numbers that no longer are needed, Hickson ditched the term "HTML5" in favor of just "HTML." And he said he'd like to see the W3C follow suit.
Don't expect the standards group to do so, though.
The W3C has always revised its standards, Jacobs said. "That doesn't mean everybody wants the nightly build of a specification," he said, referring to the software development practice of building a new test version of software every night to include programmers' latest patches. "We also have stable versions of standards, because there are some communities who need those for the level of interoperability they require...We think both innovation and stability are valuable, and they are not mutually exclusive."
Another factor is intellectual-property rights--specifically, patents. Those who participate in creating the W3C's specifications agree not to sue those implementing the specification for infringement of any patents those participants own. It's a bit of legal reassurance in a technology world that has plenty of patent risks, but technically that assurance only comes with the final version of a specification.
The final schedule
What exactly will happen between now and mid-2014 with HTML5? Several steps, according to Le Hegaret and Jacobs:
In May 2011 comes the first "last call" draft of HTML5. This version is feature-complete, meaning no new features will be added, but that existing features will be refined. The W3C expects to deal with thousands of comments through this phase, some of them likely to lead to "substantial" changes.
Likely by the end of 2011, the W3C will issue a second last-call version and begin a second round of refinements.
In the second quarter of 2012, a new phase begins, in which "implementors" of the specification--browser makers, essentially--provide feedback. During this phase, the W3C concentrates on a suite of thousands of tests to see if implementations of HTML5 really do get the same results when interpreting a Web page's code.
The culmination of this phase is a "candidate recommendation" of the HTML5 spec and at least two "interoperable implementations"--in other words, two different browsers that produce the same results on the test cases. The implementors' feedback is scheduled for completion by the first quarter of 2014.
Last comes a final review period of about six weeks, then some time to get the promotional gears engaged.
Then, in the second quarter of 2014, HTML5 should be done.
"We're excited to be able to say we now have a time frame," Jacobs said.