Zeldman has written a fair bit more about HTML5 since this article. [1] One of those articles has the title "Loving HTML5", and says "the more I study the direction HTML5 is taking, the better I like it." [2]

Along with various other well-known 'standardistas', including Dan Cederholm, Tantek Çelik and Eric Meyer, he has also formed the HTML5 Super Friends [3], an interest group to support the HTML5 development effort and push for their own concerns about the new standard [4] to be recognised.

Ian Hickson, the editor of the specification, has told Zeldman (and anyone else) to email him or one of the HTML5 mailing lists if they want anyone to pay attention to their suggestions (in other words, don't expect the working group to take any notice if you just write stuff on your blog). [5]

[1] http://www.zeldman.com/category/html5/

[2] http://www.zeldman.com/2009/08/31/loving-html5/

[3] http://www.zeldman.com/superfriends/

[4] http://www.zeldman.com/superfriends/guide/

[5] http://www.zeldman.com/2009/09/04/html5-redefines-footer/#co...

isn't this an old article that pre-dates Zeldam's involvement with the HTML 5 spec (http://www.zeldman.com/superfriends/ "We, the undersigned, wish to declare our support for the direction in which the HTML5 specification is heading").

That's correct. Now Zeldman is lovin' HTML5:

http://www.zeldman.com/2009/08/31/loving-html5/

Despite endless debate on these abstract HTML things, I'm wondering is there a better approach we could take, so that every party involved will be better off.

Like everything that grows too complex, someone will find a way to hide a lot of it behind a simple abstraction layer, and call this layer a framework. And when that framework is around long enough and has grown sufficiently complex, it will get abstracted again.

It seems to me that when things become messes, that's when people step in to clean them.

All problems in Computer Science can be solved by another layer of abstraction, except for too many layers of abstraction.

-- Butler Lampson

> And when that framework is around long enough and has grown sufficiently complex, it will get abstracted again.

Eventually, it's just turtles all the way down.

Why does the language have to be backward compatible with HTML 4? As long as the browser can handle them both, they could be completely independent, couldn't they? And since all current browsers handle the current language, all that would be needed would be code to handle the new one. just let the browser know which it is in the header.