Free COVID at-home test kits The cost of flying internationally Spider-Man: Across the Spider-Verse trailer Omicron vs. delta Walmart PS5 and Xbox Series X restock Cyber Week deals

Growing pains afflict HTML5 standardization

Emotional disagreements between two groups are disrupting the creation of the high-profile standard at the heart of the next-generation Web.

Listening to marketing messages from companies such as Apple and Google, one might think HTML5, the next-generation Web page standard, is ready to take the Net by storm.

But the words of those producing the specification show an HTML governance process that can be stormy, fractious, and far from settled down. The World Wide Web Consortium's return to HTML standardization after years of absence has produced tensions with the more informal Web Hypertext Application Working Group (WHATWG) that shouldered the HTML burden during that absence.

Some examples of language that's cropped up this month on the W3C's HTML Working Group mailing list: "childish," "intolerable," "ridiculous," "shenanigans." And there's a concrete manifestation of the divisiveness: The WHATWG and W3C versions of the HTML5 specification, though both stemming from the same source material, have diverged in some areas.

Some of the differences are relatively minor, and there are strong incentives to converge the two drafts of the HTML5 specification so that browser makers and Web developers aren't faced with the prospect of incompatibilities. In the meantime, though, the overseers of the Web are clashing during a time when their important new standard is just arriving in the spotlight.

"It's not an ideal situation. You want as much energy devoted to improving the spec and as little lost to the friction costs," said Mike Shaver, Mozilla's vice president of engineering. In the long run, though, Shaver said, it's like writing software, where improvements cause temporary disruptions: "Sometimes you add something new, you have stability problems, you clean them up before they ship."

HTML back on the front burner
HTML, the Hypertext Markup Language, is a specification that describes a Web page; the W3C released the most recent version, 4.01, an eternity ago, in 1999. When the W3C pursued an incompatible, and largely unsuccessful alternative called XHTML 2 and refused to amend HTML, browser makers Opera Software, Mozilla, and Apple formed the WHATWG to make their advancements regardless. The WHATWG's work ultimately became what's called HTML5 today, and the W3C eventually concluded it needed to be involved again in what had become the path forward for the Web.

HTML5 has a number of standout features--the ability to embed video and audio directly into a Web page without relying on Adobe Systems' Flash Player plug-in, for example. Other new HTML features, such as Canvas for 2D graphics and geolocation for Web sites to know a person's physical location, are being standardized but aren't in the W3C draft of HTML5.

Some less glamorous parts of HTML5 are important, too, including a major effort to codify how Web browsers should digest Web pages. Some of the success of HTML5 therefore will be judged by compatibility: Developers can write a single version of a Web page rather than creating different versions for different browsers.

But where will those developers look to find that standard? The W3C, a recognized standards body that includes the participation of Microsoft and carries patent policy that attempts to ease patent-infringement worries? Or the WHATWG, where many technical details and new developments are hammered out, and which houses the "sane" version of the HTML specification, according to Ian "Hixie" Hickson, who edits both versions?

Don't expect Hickson to bend his principles just so the two drafts are harmonized. "I value technical merit even higher than convergence," Hickson said Friday in one debate.

The differences between the drafts lie more on the periphery than at the core--Hickson calls them "trivial from a practical perspective." For example, one involves a "ping" feature that could help publishers track clicks on links without resorting to indirections that take visitors to an intermediate page just for tracking. Another involves "microdata," a technology for adding computer-readable tags into Web pages to describe data such as addresses; it competes with another technology called RDFa.

The differences in some cases amount to a matter of publishing specifications as separate modules rather than Hickson's preference for the more monolithic form he uses for the WHATWG version. But they're differences nonetheless. And as HTML develops, its overseers will undoubtedly have to tackle bigger matters than packaging.

The divergent W3C and WHATWG specifications has some concerned. "The whole point of an open technical standard is to promote interoperability between implementations, and having conflicting or ambiguous specs will not result in that goal," said Doug Schepers, who works on specifications such as Scalable Vector Graphics (SVG) for the W3C, in a Friday WHATWG mailing list message calling for technical parity and suggesting the WHATWG appoint a W3C liaison "with a reputation for reasonable neutrality."

Tempers flare
Hickson, a WHATWG charter member who's been the lead HTML5 editor since 2003, helped launch the recent bout of fractiousness. Though he's a Google employee, he acts independently, and he spoke his mind freely when criticizing the W3C HTML Working Group's rationale for deciding to remove various elements of HTML5:

"These are the most wishy-washy, noncommittal, incomprehensible, and nongeneral arguments I have ever seen in my specification editing career," Hickson said in an e-mail.

That characterization didn't sit well with Sam Ruby, an IBM employee who, along with Apple's Maciej Stachowiak and Microsoft's Paul Cotton, are the co-chairs of the W3C's HTML Working Group. The three hold responsibility for the W3C's recommendations for the HTML specification, and Ruby countered with criticisms of Hickson's reasons for action.

"Items appear in the WHATWG specification based on the individual discretion of the author [Hickson], often without any documented rationale," Ruby said, defending the W3C's choices as documented.

Hickson, evidently unconvinced, changed the WHATWG version of the HTML5 draft, saying it included a particular paragraph that the W3C version omitted "for political reasons." In an e-mail, Hickson said, "...from a technical point of view, your decisions are all arbitrary. You put [forward] pseudo-technical reasoning for your decisions, then dogmatically reference them whenever anyone asks you to explain them...The WHATWG draft continues to exist because it's the only way to have a specification that actually makes sense in the face of the ridiculous decisions you keep making."

Ruby, with the agreement of Cotton, fired back that the move to include the language was "intolerable."

"As long as the WHATWG draft is used as a vehicle to discredit the W3C, then we will insist that all references to the WHATWG specification be removed from the W3C specification," Ruby said. Eventually Hickson relented on the "political" wording.

The tensions haven't dissipated, though. Ruby and Hickson clashed on still another matter Friday and Saturday, newer dispute over a matter concerning browser plug-ins.

Politics versus pragmatism
And Hickson hasn't changed his views of the W3C decisions.

"We shouldn't let the near-term politics of the Web affect the long-term technical quality of the Web," he said in an interview. "The way to get the drafts to converge again would be for the W3C to change from a model of appeasement, where the most vocal objectors get to push through their agenda, to a model based purely on the technical merits of proposals, where the volume of objections is simply ignored and only the value of what is being suggested matters."

Ruby, Stochowiak, and Cotton didn't respond to a request for comment. But Philippe Le Hegaret, who leads the W3C's work for Web standards including HTML5, Cascading Style Sheets, and Scalable Vector Graphics, took a conciliatory tone.

"The Working Group is making the decision about what gets in and what gets out," he said, though in practice it's more complicated because "we are using the editor [Hickson] as a filter," a first step of assessing whether a change is good or bad. Those who aren't happy with Hickson's decisions can appeal them to the Working Group, Le Hegaret said.

And while some have called for the W3C to sever its ties with the WHATWG, or at least to excise its specification's references to the WHATWG's, the reality is that many people participate in both groups. "We don't want to pretend we're the only group of innovators," Le Hegaret said.

For many participants, "politics" isn't such a dirty word but rather the reality that nontechnical factors are inevitable when getting business done. In other words, a pragmatic ethos as well as technical idealism can serve the Web.

"There are many, many considerations--timing, business cycles, product cycles," said Ian Jacobs, head of communications for the W3C.

"I think all decisions are made for political reasons. Mozilla has an agenda. IBM has an agenda. Microsoft has an agenda. Ian [Hickson] has an agenda. Google has an agenda," said Mozilla's Shaver, though he added that he'd prefer more transparency in hashing the issues out. "The value of having Microsoft implement some stuff is worth making some of these trades."

Browsers on board
Indeed, Microsoft is a key player. Long the whipping boy of the browser industry for releasing versions of IE that ran roughshod over Web standards, the company is trying to turn over a new leaf, including an effort to participate actively in HTML development at the W3C. Microsoft's involvement is very important--it remains the dominant browser maker despite gradual declines--but the company has no interest in the WHATWG.

"Microsoft has participated for many years in the W3C and continues to be a strong believer that interoperability is best achieved through a collaborative standards process--a formal process, long exemplified by the W3C, that is open to all stakeholders and governed by a transparent and delineated set of rules," the company said in a statement.

Apple didn't comment, but Google said standardization is about resolving disagreements. "Google is committed to making the Web platform more powerful by working with standards bodies to introduce new capabilities and better define existing capabilities. Open debate is an important part of the standardization process as it ensures we reach a specification that all parties can agree to," the company said.

And Opera--along with browser pioneer Netscape, one of Hickson's former employers--said Web standards need more than the WHATWG. "While WHATWG allowed a few companies to start high-quality development, it isn't so well suited to actually creating a global standard along with the stakeholder community necessary to do that well," said Charles McCathieNevile of Opera's standards group.

The HTML5 specification, while not done, is mature enough in at least some areas for real-world use. Firefox 4 is being fitted right now with an HTML5 parser to process Web pages, and the latest developer version of Chrome takes the first of two steps toward its own HTML5 parser, according to Google programmer Adam Barth, through developments in the WebKit browser project that Apple also uses.

Reins of power
The way Hickson sees it, the browser makers--the "implementors" of the standard, in standards-speak--get the final say.

"In practice I have no power whatsoever, even in the WHATWG: I am totally at the mercy of implementors," he said, adding that the same is true of the W3C's HTML work. "This is the sad reality of spec writing: if the implementors disagree with what the spec says, they'll do what they want, and the spec becomes just a very dry form of science fiction. My role is basically to take the input of everyone, then try to convince the browser vendors to all agree on one particular direction." Perhaps it's no coincidence his e-mail signature reads, "Things that are impossible just take longer."

Hickson has garnered plenty of respect as he's tackled Herculean tasks--for one example, see his 23,000-word attempt to assess and respond to more than 620 e-mails concerning mathematical text and Scalable Vector Graphics in HTML. But he's also raised hackles. One critic is Shelley Powers, a former member of the W3C's HTML Working Group and author or co-author of 18 books, mostly about Web technologies.

"Ian is a very capable person, but he believes that the best, quickest way to create a specification is if one person makes all the decisions about what is in a specification. Ian also has strong, rather inflexible, opinions about the future of the Web, and the future of HTML. One person with such strong, inflexible opinions, and with virtually unlimited control over the HTML5 specification contents, is not a good thing," Powers said.

"In my opinion, the HTML5 specification would be vastly improved if there were a small team of experienced specification editors, rather than one single person given such unlimited control," she said, and she thinks the existence of the WHATWG today "does more harm than good."

To critics who say he has too much power, Hickson said they're welcome to try their own hand:

"They have as much power as I do," Hickson said. "Seriously--the spec is licensed under a completely open license, anyone can take it and edit it and have the same level of 'control' as I do. All they have to do is do a better job in the eyes of the browser vendors."

W3C ambitions
The HTML disputes come at a time when the W3C, under the leadership of new chief executive Jeff Jaffe, is trying to reclaim some of its power.

"There is much new innovation, and the Web will benefit if the community brings their work to W3C," Jaffe said last week in a blog post, adding that the W3C is trying to become more agile and open.

In the eyes of Julian Reschke, a consultant at Greenbytes and one of the editors of the Hypertext Transfer Protocol specification, some of the present difficulties are the result of that reassertion of the W3C's power.

"The W3C working group tries to make decisions, and by doing so occasionally overrules Ian Hickson," Reschke said. "Each time it's a fight, because he doesn't like to be overruled."

Expect more tussling in the future. The W3C pushed some specifications out of HTML5 for later consideration, and those technologies will surface again in the standards process. And as the Web's importance and sophistication continues to grow, so will the crowd of people who want to influence it.

Thus, the repercussions of June's HTML disagreements could extend well into the future.

One possibility is a growing rift between two camps that are both deeply involved in the standard. But given the players' commitment and their continuing work together, a rosier scenario is perhaps more likely: oversight of the future Web isn't collapsing, it's just rebalancing.