What does HTML5 mean to you?

HTML5 is used in two ways nowadays. One is specifically within W3C, where it’s one specification, one specific document. Nowadays in the larger audience, it’s more used to mean what we call ourselves the “open web platform” — a set of specifications that are all interacting together, including HTML5, CSS3, SVG, the geolocation API. People use the term HTML5 to refer to this big platform, but we’re talking about a string of specifications that are not developing at the same speed or being rolled out at the same speed.

We’re moving from a document-centric web to a more application-centric web. In order to deploy your application, it’s going to be done using HTML, JavaScript and CSS. Sometimes you’ll know it because it runs in the browser. Sometimes you won’t know because it’s not in the browser: a good example is iTunes — it’s HTML, CSS and Javascript underneath. Users don’t know — it’s an application from their point of view — but it’s based on web technologies.

Why is HTML5 necessary?

To a certain extent we always have this dream of writing an application once and have it being able to run everywhere. With HTML5 we feel like we’re developing that: HTML5 is being pushed as one of the solutions to answer this question. This time not only is it on every single desktop, but mobile too. It’s becoming easier and easier to adapt your application depending on the size of the screen. This idea of writing once and running anywhere is important.

In the past, web apps were limited in what they were able to do, we are starting to expand it in many directions.

How much of the specification is codifying what’s out there already, and how much is laying a path for the future?

The reason people are talking about HTML5 today is not because we’re codifying things, it’s because of the new features like videos and canvas. I think we’re striking a good balance, but it’s not clear whether there’s a real interest in implementing the more obscure features, like datagrid.

It will still take some time for those technologies to be found on every website, and that’s because we cannot upgrade the entire web and every browser in one go. Today we still have IE6 out there, IE7 doesn’t support anything from HTML5, IE8 supports a little bit and IE9 is doing much better, but before we can get that deployed it’s going to take several years. For the web developer in the street, if they want to deploy HTML5 they can, but they’re going to need a fallback mechanism.

One of the reasons, and let’s be honest, was because one of the vendors did not upgrade their browser for several years. If you try to deploy the Canvas element today, and try to use it, it will work all the browsers that are deployed except one. That’s definitely had an impact in trying to standardise the web. The fact that they’ve decided to do it with their upcoming version, that’s great… we’re no longer concerned about the deployment of HTML5.

The W3C has come in for a lot of criticism for its original plan to transition the web to XHTML. What’s your opinion on that?

While we’ve put the idea of a new version of XHTML to rest, it may resurface in a few years.

Our problem was to deploy XML on the web. It’s used in plenty of cases, but the one area where it wasn’t fully used was HTML.

And what changed when you decided to end development of XHTML in favor of HTML?

I think the effect that it had was that we stopped wondering what solution to push for. For two years we were pushing both solutions. The electronic books industry, they were the ones who were heavily relying on XHTML and were disappointed when we announced last year in June that we were stopping development of XHTML2 — but they’re now starting to move to HTML5.

Why has it taken so long to finalise the spec?

The bulk of the work we’re currently doing, and the reason it’s taking so long, is that we’re working on accessibility issues: you can build a great 2D game with Canvas, but it’s totally inaccessible. People will have to be careful when they use these technologies, but that’s always been the case. We are certainly more conscientious in terms of implementation. One of the things we are already doing is testing: this wasn’t done for HTML4 when it was released 10 years ago. We need to make sure that people aren’t just agreeing on functionality just for the sake of making the document move forward, but without commitment to implementation.

What about HTML6 — is it going to happen?

We certainly intend to make revisions to the HTML5 spec — we haven’t started working on HTML6, but we intend to start on it before the end of next year. From our point of view, HTML5 is not the final version of HTML, we’ll keep adding new features as they become important. The thing we’re facing right now is that there’s so much happening. The bulk of the innovation is happening at the API level… and the browser vendors, they need to digest all of that. At the moment I do not see a lot of push to come up with HTML6 right away.