HTML5 to be complete in 2014

html specification dropping version numbers html5 logo

Ask any person first stepping into the world of HTML whether the markup language used for the Web is easy and straightforward, you’ll probably get a variety of answers. For developers, programmers, and folks used to dealing with technical specifications, HTML is reasonably well-documented: there are heaps of books, tutorials, mailing lists, online resources, and (of course) millions of Web sites from which to glean information. However, ask anyone with a more prosaic background whether HTML is straightforward and the answer is likely a resounding “no,” with a mess of completing (and often conflicting) set of standards, acronyms, and versions clamoring for attention.

The W3C tried to put a pretty patina over the technology mess by promoting a collection of Web standards under the label HTML5, only to have the Web Hypertext Application Working Group (WHATWG) say that HTML5 would be a constantly-changing “living document” with no fixed version numbers. And now the W3C’s HTML working group is muddying the waters still further, saying there will indeed be a finalized version of HTML5…it’s just not due until the middle of 2014.

Although the 2014 completion date for HTML5 seems like a long way off, the W3C plans to release a “last call” version of the specification in the second quarter of 2011. The LC spec will include ever feature the W3C believes should be included in the HTMl5 spec: although those features will be refined and changed as the spec is refined, the W3C does not expect to add any new features to the spec after the LC release. That should give Web developers some guidance on how to architect and develop sites and services.

However, the W3C’s last-call and final spec may or may not have any impact on the WHATWG’s working document—although both the WHATWG’s and W3C’s versions of HTML5 are maintained by Google employee Ian Hickson, there are already differences between the two versions, and the versions of HTML5 may increasingly diverge once the W3C’s Last Call spec is ruled “feature complete.”

Part of the reason the W3C’s official process takes so long is that it must coordinate input from a broad group of stakeholders. Some of that coordination is rights clearance: if any aspect of the HTML spec is covered by a patent, the W3C must secure an agreement from the patent holder now to sue anyone implementing the HTML specification—and that safety from being sued only comes with a finalized specification. The W3C also creates implementers versions of the specification in order to iron out details in actually getting HTML to work; after that, there have to be at least two “interoperable” implementations of the HTML spec—basically, two compliant browsers—before the spec can be finalized.