You can read some of their thoughts in the piece itself, but — as is always the way with reporting — there was a vast amount of material that fell by the wayside for one reason or another. However, the interviews I did were extremely interesting, and informative… not simply because they tell you about the technology itself, but also reveal some interesting information about the politics of the web and the way things might be going.
It’s a bit of a departure for this blog, but since I didn’t keep a public notebook on the article itself, I thought it might be worth publishing their thoughts here. I’m going to put up one a day for about a week.
First up is Ian Hickson, the editor of WHATWG — the splinter group that formed in the wake of a split between browser vendors and the World Wide Web Consortium. It’s really been the force that has pushed HTML5 to the point today where Google can launch an operating system using it, and Steve Jobs can claim it might give Flash a run for its money. As it happens, Ian works for Google these days.
Here’s a lightly edited transcript of what he told me:
On the next version of HTML — and whether there will be one
“From the point of view of what I work on, HTML5 is old hat now! The WHATWG is already moving on to the next generation of HTML, and one of the changes we’re making is that we’re no longer numbering the versions… in practice it hasn’t made much sense. Browsers started implementing HTML”5” stuff before all the HTML”4” stuff was implemented, and they ship whatever they have at the time they ship… it’s not like the version numbers really mean anything useful. So there’s just “HTML”, and we keep improving it.”
“I don’t think we should have a new version — I think we should just continually maintain the HTML specification (and all the other Web technologies), adding features in sync with browsers, removing stuff that has failed in the marketplace, and so on. The model of coming down from the ivory tower with a new spec written on stone tablets and handing them out and then going back into the tower for a decade really is no longer tenable in the market today. We have to treat the specs just like we treat software, as living documents. Once we stop working on a specification, once we no longer fix bugs in a specification, that technology might as well be dead — or at least, the specification will become irrelevant.”
On the most interesting aspects of HTML5
“I could provide a litany of features we’ve added to HTML since the WHATWG started working on it — from the well-known <video>, <canvas>, the new form controls, microdata, <figure>, <progress>, the new section elements… to the smaller, less visible stuff like scoped <style> blocks, <a> around flow content, async scripts, reversed lists… to the newer experimental stuff that’s still a work in progress, like <device>, WebSRT, <command>.”
“But I come from a browser QA background; I’ve literally spent years writing tests to check every last bit of a spec. So the main thing that I like about the new HTML spec is that it finally defines so many things that were basically just folklore before. For example, it is the first spec to define the Window object, or the event loop mechanism, or document.write(), or how to parse HTML and handle errors. This is going to really help get browsers to have fewer differences, which should really help Web developers.”
On editing the specification
“Being an editor is usually fun — people provide feedback in the form of use cases, data, research, arguments for and against features, and then I take all the feedback on a topic into account, reply to all the emails on the topic, and come up with a proposal to address their requests. Sometimes it works better than others. So long as all the browser vendors agree, my life is easy — I just spec what makes them happy. Essentially, they have all the power; I’m only able to maneuver within the areas that they don’t care much about. If they disagree, I have to go to each one and see if they’re willing to converge on something.”
“Sometimes, I just can’t get all the browser vendors to agree on something, and then it’s time to cut my losses and move on to something else. This happened, for instance, with the Web SQL Database specification, or with the discussion of which video codec the <video> element should use. Luckily there’s still plenty of low-hanging fruit where we can get real productive work done easily.”
On adding and removing elements from the HTML specification
“There are various criteria we take into account when deciding what should be in and what should not — how much legacy content uses it, how many browsers are willing to implement the feature, how many already implement the feature, how many are willing to remove their implementations of the feature, whether the feature really solves any real problems.”
On the split with the W3C
“The W3C lost sight of the fact that they have no power. That’s really all there is to it. Anyone can write a specification, but if nobody implements it, what is it but a particularly dry form of science fiction?”
“Around the time we were first trying to get the W3C to work on HTML, I often heard things from W3C people along the lines of “browsers are dead”, with the implication that new standards wouldn’t need implementations.”
“This has never made any sense to me. I guess they eventually realised that browsers actually might have some importance to their work.”