W3C home > Mailing lists > Public > public-html@w3.org > April 2007

Re: legacy of incompetence? [was: a compromise to the versioning debate]

From: Henk-Jan de Boer <html-wg@hjdeboer.nl>
Date: Wed, 18 Apr 2007 01:13:23 +0200
Message-ID: <46255493.6060502@hjdeboer.nl>
To: Dão Gottwald <dao@design-noir.de>, HTML WG <public-html@w3.org>

Dão Gottwald schreef op 17-4-2007 16:44:
> Nobody can prevent MS from doing one more mode switch for <!DOCTYPE 
> html>. And this single switch wouldn't be a problem, in my opinion. 
> Other browsers don't have to follow that route, as their standards 
> modes are incompatible with IE even today. But AFAICT, Microsoft's 
> argument is that HTML>5 won't be necessarily backwards-compatible and 
> could require a doctype switch (hence versioning). That doesn't sound 
> like a valid argument to me, given that we all want future HTML 
> versions to be backwards-compatible and that existing switches aren't 
> there to represent HTML versions but to emulate the wrong rendering 
> behaviour from older browsers. Now I fear that MS actually wants more 
> switches because their initial implementation for the new HTML5 
> rendering mode (that includes CSS and DOM at least) could be screwed 
> up again. I don't think that's acceptable.

We're on the same track here. The only solution I see to that problem is 
when Microsoft would release bugfixes to mend broken implementations 
_very frequently_ so that web authors don't even get time to get used to 
and rely on those bugs. But then I guess that would be an unrealistic 
development path for a browser vendor with a user base that big?

>
>> Wouldn't it be better to pick up the versioning debate later on, when 
>> Apple, Opera and Mozilla can claim that HTML5 is fully backwards 
>> compatible with existing content?
>
> IE, due to its broken standards mode, often gets different content 
> served than Safari, Opera and Firefox. Authors expect IE to be broken, 
> hence the risk to break sites when fixing bugs. You can't test this 
> with other browsers.

Thanks, I did a wrong assumption when I wrote that. You're very right; 
but here's that paradox again: how can HTML5 claim to be fully backwards 
compatible with all legacy content, when a lot of that content is 
written, based on incompatibilities of the past?

Kind regards,
H.J. de Boer
Received on Tuesday, 17 April 2007 23:12:04 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:15:53 GMT