Re: Versioning and html[5]

On 13 Apr 2007, at 20:07, Chris Wilson wrote:

>> Supporting
>> it, in all its versions, is prohibitively expensive for anyone but
>> Microsoft.
>>
>
> Supporting all Office formats from all past versions is incredibly  
> expensive, period.  Microsoft just already paid some of that cost.
>

And for me, that underlines the point of this non-versioned HTML  
concept. If you version HTML then you're imposing similar chaos onto  
a future UA development as a new word processor has with DOC and DOCX  
now. Yes, HTML has more specs than DOC but the quantity if  
unspecified behaviour will only increase with time if the document is  
versioned rather than fixed.

As Ian says, people are writing content for the web to last forever.  
Of course that sounds grand, but having one, backward interoperable  
spec is the best way to serve content publishers.

I have a question for Chris Wilson and other versioning advocates at  
this point:

What if you're right? What if actually it is unsustainable to develop  
HTML in an evolutionary manner beyond HTML5? Maybe Ian Hickson and  
the WHATWG, bright as they are, have missed some critical facet that  
means we do need a discrete switch between revisions after all?

What if you're right? How does that block this group from  
_attempting_ to build a better, evolutionary and backward compatible  
HTML? Why can we not approach this first ‘new’ HTML with the complete  
and committed intent that it be evolutionary over HTML4 and that  
HTML6 will be evolutionary over HTML5? Why can't we try it this —  
better — way?

You'll still get a new DOCTYPE — <!DOCTYPE html> — so IEs legacy  
features can be switched one last time if you guys really need it.  
And then, we try to make this work.

In 5 years time people will think about a second iteration of the  
HTML-WG process;  alterations and enhancements to HTML5. If you're  
right and HTML does need a version flag to be implementable, what  
stops us looking at this question again? If versioning really is  
required then to move from ‘HTML5’ to ‘HTML6’ you can provide  
evidence and argument based on experience of this model. And if  
you're right, then we can add a version flag in HTML6.

I don't understand why we can't pursue the best development process  
for HTML — non-versioned, evolutionary — and come back to versions at  
HTML6 if for some reason it won't work. I think I'm missing the part  
where this is actually a problem _right now_?

Regards,

Ben

(Apologies to Chris Wilson, who I accidentally sent this message to  
directly earlier this morning)

Received on Saturday, 14 April 2007 11:41:43 UTC