Tag Archives: what wg

The two specifications of HTML 5.0: WHAT WG vs W3C

I’m just back from a workshop on HTML 5, led by web standards advocate and CSS expert Molly Holzschlag. It proved an illuminating session, though not quite in the way I had expected. Holzschlag, who works for Opera, was keen to convey the ideology behind HTML 5 rather than giving us a blow-by-blow tour of its features (though she did a little of that). She was also open about its problems, explaining that the spec is in flux and everything may change – “we make it up as we go along” – and talking about the politics as well as the technical aspects.

In her view, Microsoft is now fully on-board with IE, and committed to implementing the W3C HTML spec as it evolves. So too are Mozilla and Opera. She is less warm in this respect towards Apple, Google and Adobe, who she described as the “new Microsoft”, meaning I think that their business interests may be detrimental to their work on progressing the standard.

It is surprising to see Google mentioned in this context, since it is the company most obviously concerned to advance the browser’s capabilities, thus increasing the capabilities of its web-based platform. Ian Hickson, who is the editor of the HTML 5 specification, works for Google. Still, HTML 5 is a subject full of contradictions. One of its most curious aspects is that there are two HTML 5 specifications, one at WHAT WG, and one at the W3C, both edited by Hickson.

The history is that at one time the W3C, the official body in charge of the HTML specification, decided to replace HTML with a stricter XML-compliant language called XHTML. Real-world adoption was limited, and WHAT WG was set up by Google and browser vendors as a renegade group to work on a new version of HTML outside the W3C. When it became clear that XHTML was not achieving its goals, and that HTML 5.0 was meeting real needs, the W3C changed direction, stopped working on XHTML, and adopted the WHAT WG spec.

At this point you would have thought that WHAT WG might have closed itself down, job done. That is not the case though; it continues to work on its own version of the spec. I asked Holzschlag why this is, given that the existence of two HTML 5 specifications seems on the face of it to be destructive:

I think it’s very destructive. It’s very problematic. The WHAT working group is into innovation, and pushing the envelope, where they can’t do that in the W3C. The reason why the W3C’s stuff is important … is because it’s about open standards. The WHAT working group has no validation or validity or standing as an organisation other than its own self-involvement. The W3C is clearly the authority for most of these things.

Holzschlag emphasized that the W3C is a cross-industry body, with every browser maker and other interested parties such as Adobe represented.

Get us all round the table, and once we’ve spilt enough blood [laughs] we get on with the work and that actually goes through a very rigorous process, which a lot of people criticise and I feel it could be streamlined as well, but the bottom line is to ensure that it stays open, and it’s open standards, whereas the WHAT working group can decide any day that they want to close that door. At the W3C that can’t happen. That’s why if you’re really going to commit to anything in HTML 5, go with the W3C specs not with WHAT WG.

It’s a political issue in part, and in part it’s an ego issue. I think that Ian and his mates are great, very bright people but they are not totally mature yet … and I think that there’s a sense of self-importance going on, to be perfectly honest … I’m a little concerned about the monoculture that HTML 5 has created. So that exists and is a known factor. Everything I’ve said is nothing that hasn’t been said before publicly.

Strong words; yet overall Holzschlag conveys great enthusiasm for HTML 5 and its potential. She says that the mere fact of having all the leading browser vendors on board and talking to one another is of great significance.

But does HTML 5 exist? In some ways it does not; it is work in progress and not implemented consistently across browsers yet. That said, Holzschlag noted that the latest versions of the main browsers already implement significant parts of HTML 5; we will no doubt see more of it in Internet Explorer 9, for example. Even though Hickson said HTML 5 might not be done until 2022, it will be usable long before that.