'

What Google thinks of HTML 5: a buzzword?

Thinking about WHATWG's rather Orwellian approach to the specification formerly known as HTML 5, where the 'specification' becomes a 'living...

Thinking about WHATWG's rather Orwellian approach to the specification formerly known as HTML 5, where the 'specification' becomes a 'living document' that will likely have changed by the time a developer has finished implementing the version they last looked at (and wondering whether the resurgence of WHATWG after the recent focus on the official Web standard body, W3C, is a reaction to how much more involved Microsoft is with the W3C now and how much of the W3C HTML 5/SVG/CSS 3 specification will be in IE 9), I was reminded of the way that Ian Fette, a product manager on the Google Chrome team, described HTML 5 last year at Google IO. (Although Google wasn't one of the founder members of WHATWG, the current editor of WHATWG is Google's Ian Hickson, author of the controversial Acid 3 test.)

"A lot of people use HTML 5 to mean the new capabilities added since roughly 2008," he pointed out. "Is that technically correct? No." But he excused the imprecisions, ironically because it's been difficult to keep up with the way HTML 5 has changed (for example, when originally proposed, the HTML 5 video tag would have specified a particular codec; W3C couldn't agree on what codec it would be and left it up to the browser to specific why codec its video tag supports - an issue that Google has rather high-handedly tried to force by only supporting WebM and Theora in Chrome). Or as he put it, "a lot of things have gone into and come out of HTML 5 in the specification sense."

So in talking about the progress Chrome has made in the previous year, he declared; "When I use HTML 5, I'm going to use it in the buzzword sense."

The Web is an evolving platform, and there is always going to be a tension between developers who want to use the latest and greatest technologies to do something new and exciting that the official specification doesn’t yet enable, and the business developers who want to build something to a specification they can expect to be supported for a decade because that's how long they're going to have to support it for. Not everyone wants to update their browser every night or even every week.

There's also a tension between the engineering, test and specification-based approach Microsoft is comfortable with - and has already proposed to the W3C for HTML 6 - and the 'make your case and convince the editor and the invited members of WHATWG' benevolent dictatorship route. A healthy tension between innovation, adoption, refinement and standardisation keeps the Web moving along for the developers who want to be on the bleeding edge without leaving those who need consistency too far behind. But for that to carry on working, a point in time specification like HTML 5 needs to be a bit more than a buzzword. If people keep using the term HTML 5 incorrectly (referring to 'HTML 5' test sites that include non HTML 5 features, lumping in any new Web technology no matter what its maturity or codification or just making it mean what they want it to mean, Humpty Dumpty style) the solution is not to file the version number off the HTML spec and make it harder to know exactly what Web technologies a tool is claiming to support. Surely it would be better to educate people about what HTML 5 actually covers, refer to it consistently and - oh, I don’t know - have a logo that you can use on products that implement it?

Because if HTML 5 is just a buzzword then we’re back to browser makers declaring that 'my browser is better than yours because I say so'.

Mary Branscombe