W3C names four new editors of HTML5 spec

In some ways, I quite like the idea of running two versions... you've basically then got the choice of HTML, for the want of a better word, Experimental... and then what should be a more supported standard, where you know what to expect and browsers can target specific snapshots.

All the vendors are really gunning for the support, so it's fairly likely the snapshots would be well supported and browsers with faster refresh cycles would also be able to add some bleeding edge features. Since MS doesn't really add support to a browser version once it's shipped it also gives a target point for IE, a rolling standard was never really a good thing for Microsoft as much as I now value their efforts to follow standards.

Crazy

As I understand it the HTML5 specification is running about 4 years behind schedule.

Isn't it about time someone pulled their finger out?

Just mirroring what devs already do.

CSS3 colours have been in common use for years now, even thought CSS3 is still under development - any transparency on a website other than from GIF/PNG is thanks to this. Likewise border-radius for round borders, CSS3 transform (think google maps and openstreetmap), Server-Sent events (gaining traction for webapps), gradients video and other fun stuff.

All of these can be applied now, across browsers, without too much difficulty (I'm not too familiar with IE but I believe IE9 is not too bad, astonishingly). There is a steering commitee of sorts and the architects of the various browsers collaberate: things are not like the bad old days, this stuff is documented and tested so there's no need to guess and the vast majority of browser hacks are no longer necessary - usually it's just a matter of browser-dependent CSS prefixes, becuase the official spec is unfinalized.

Meanwhile the W3C, with the same goals and even mostly the same people, only managed to ratify CSS2.1 last June. It's prototype-led development vs. mythical-man-month style formal specification. Without the WHATWG we'd have to rely on just the latter and that, IMHO, would be a disaster.

ECMAScript, 4th Edition all over again.

Business

The reason the browser makers are so keen on the new model rather than the W3C way of keeping snapshots that last a decade is that it prevents any one browser becoming the defacto standard, like IE used to be. And it stops old versions lingering on for years after they should have disappeared.

There's pressure on IE, Firefox, Chrome and Opera to keep moving and bring their userbases with them. That's why most browsers now have silly version numbers and silent updates. Version numbers don't matter anymore, you generally need to have the latest version of a browser otherwise websites will start breaking within a couple of years.

And unlike the good old days, there are rarely fallbacks for Javascript/DOM, CSS, plugin content etc. Still, I suppose when you look back at how getting suitable alt text on images was like pulling teeth back in the 90s, no wonder there's so few sites that still work without their fancy AJAX frameworks. Or even just work with screenreaders.

ISO HTML Reborn

Is this the modern day version of W3C HTML - the living standard of its day. And ISO HTML, which on paper everyone could rely on because it was the 'official' rock-solid specification you could deploy and reply upon for years?

Only... I seem to recall everyone ignoring ISO and just focussing on the W3C efforts, inbetween buying into <blink>, <marquee> and document.all

OK, I know document.all isn't HTML, but you know what I mean. In other words, there'll be a couple of early IBM and Sun Microsystems solutions rolled out for a few top-level corporate clients' internal systems using W3C HTML, whilst the rest of the world ignores the W3C and uses the other specs. Later on, IBM will decide to move away from W3C HTML because it's not what the rest of the world use, and it costs too much to make their own frameworks and tools for the W3C spec rather than cheap open source solutions for the WHATWG spec which probably seamlessly handle implementation incompatibilities in the background.

The problem with standards....

is that there are too many to choose from