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

RE: Versioning and html[5]

From: Ian Hickson <ian@hixie.ch>
Date: Mon, 16 Apr 2007 22:51:40 +0000 (UTC)
To: Chris Wilson <Chris.Wilson@microsoft.com>
Cc: "public-html@w3.org" <public-html@w3.org>
Message-ID: <Pine.LNX.4.62.0704162212220.17772@dhalsim.dreamhost.com>


We seem to have wandered from the point of this thread.

I'm not sure that anything remains that we can convince each other of. You 
are convinced that you need to introduce rendering version modes to IE. I 
am convinced that those modes shouldn't be condoned by the spec (though if 
you do go down this route, I may eventually have to specify them, if other 
browsers find that they have to implement them).

Currently we haven't got an editor in this working group, nor a spec, so 
nothing can happen here anyway. However, I've taken your feedback into 
account along with everyone else's for the purposes of the HTML5 spec in 
the WHATWG. Currently the overwhelming body of argument is in favour of 
not having any explicit versioning in the spec (even if you only consider 
feedback from browser vendors, but also if you consider feedback from the 
Web authors who have spoken).

Your requirement -- to have a way to identify HTML5 documents -- is 
possible with the spec as written today, so hopefully at least for the 
near future your needs are met. From the WHATWG standpoint, should you 
need another flag in the future, or should new information come to light, 
then that can be addressed at that time.

Note that this is not any sort of attempt to close the issue for the HTML 
working group. I'm just letting you know that your feedback was taken into 
account for the WHATWG. I am not trying to imply any sort of decision was 
made for the HTML working group. As the chairmen, I'll let you and Dan 
decide how or when to close issues here, and how to move forward with 
decisions. This is not any sort of attempt to close the issue here.


On Mon, 16 Apr 2007, Chris Wilson wrote:
> >>
> >> It's really not a big deal right now.  But in HTML6, when it turns 
> >> out we forgot to normatively specify some behavior in IE5, and 
> >> Firefox and IE do it differently, we're not going to break that 
> >> behavior - meaning knowing that we're looking at HTML6 will be a good 
> >> thing.
> >
> >We've never, up to this point, needed to know what version we were 
> >looking at, and even now 17 years into the Web's development, Microsoft 
> >are the only vendor saying that there is a need.
> 
> Umm, you've HEARD of quirksmode, right?

Quirks mode is independent of version. It and standards mode both apply to 
HTML4, the mode switch is not a version switch.


> Cargo-cult boilerplate.  I like it.  It's a great descriptor for what 
> DOCTYPE is today, and what it will be in the future, even in your world.  
> I didn't invent it.  I'm not even the only one saying it should be 
> perpetuated, despite how you are trying to paint it - even <!DOCTYPE 
> html> would be cargo-cult boilerplate to get you in "one true standards 
> mode" across all browsers, in your world.

Indeed. I'd like to remove even that, but we can't, because it is required 
to trigger standards mode in legacy browsers. This is exactly one of the 
reasons why I don't want to introduce yet further mode switches.


> >The falacy is that not including the attributes somehow prevents you 
> >from implementing them.
> 
> That's known as proprietary functionality, and we get slapped for 
> implementing it.

The more important "proprietary" part here is ActiveX.


> <object> was captured in a standard.

But IE's implementation wasn't captured. IE's implementation of 
codebase="" is proprietary whether or not the attribute is in the spec as 
defined by HTML4.

However, as I've already mentioned, I'm happy to specify the behaviour in 
the HTML5 spec (though I will need help to specify the binary-level parts 
for ActiveX, as that doesn't seem documented anywhere).


> I've said (I think) that browsers won't get it perfectly right the first 
> time.  I'd love to exactly implement the spec.

Ok. Great!


> >> Not true.  <table> and <form> are still occasionally interleaved.
> >
> >The HTML5 parser spec handles this case, even with a tree structure.
> 
> It's not possible to "handle" this case - that is, have the same 
> behavior IE4 has - without a non-strict-tree structure somewhere.  
> Maybe it's in your form topology rather than your tree - I'm not sure 
> why that's better.

Because the latter doesn't violate core DOM assumptions.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'
Received on Monday, 16 April 2007 22:51:59 UTC

This archive was generated by hypermail 2.3.1 : Monday, 29 September 2014 09:38:42 UTC