CSS3 box-sizing property (resend with smaller patch).

[Germain Garand]

Le Jeudi 04 Novembre 2004 14:23, Allan Sandfeld Jensen a écrit :
> Hi!
>

Hi Allan,

> Just wanted to get peoples opinion before commiting my next patch that
> implements the suggested CSS3 property box-sizing.
> It already exists in MacIE and Opera, and in Mozilla under -moz-box-sizing,
> but it seems it is only luke-warm for CSS3.
> It is used for manually telling which box-model the webpage is using. The
> old WinIE (border-box) or the CSS standard one (content-box).
>
> Since there are already webpages out there using it, I suggest implementing
> it and under the box-sizing property. Another option would be using
> -khtml-box-sizing, but then it wouldn't work with any existing webpages.

are you aware CSS3 is a working draft, i.e, not even a proposed 
recommendation? I think it is pretty irresponsible to directly expose 
properties defined in there. They are definetly not suggested by the working 
group yet. They very well might be dead by tomorrow, for all I know.

In this specific case the blurb on top of section 9 at 
http://www.w3.org/TR/css3-box/, is clearly not comforting the viability of 
this property:

"9. The 'box-width' and 'box-height' properties (or 'box-sizing'?)
[...]
The property 'box-sizing' was first proposed to provide an upgrade path for 
certain browsers that interpreted 'width' the wrong way. 'Box-width' and 
'box-height' are proposed as improved versions of 'box-sizing'. However, 
newer versions of those browsers have already fixed the bug and it is not 
clear that these properties are really needed anymore."

>
> I discovered another odd thing while doing this. It seems that WinIE 6.0
> and Opera uses the border-box model in quirk mode, while Mozilla and
> Konqueror always uses the standard. I thought we usually tried to mimick
> WinIE behavior rather than Netscape, should we add box-sizing to the quirk
> css, now that it is easy?

MSIE is merely paying for its past mistake about the box model- it had no 
choice but to implement that backward compatible hack. 

I add the same dilemma about IE quirky font-sizes and eventually dropped the 
idea when thinking that:
1) Gecko is gaining momentum, thus turning IE quirks less and less relevants
2) we have "like Gecko" in the UA (mmh...)
3) we are even considering dropping document.all
4) it just feels wrong to inflicts to our users the same legacy contorsions 
and pains than MSIE inflicts to theirs when they want to write simple, non 
doctyped, html...

Whatever we decide, I think we must get over the IE/Gecko identity crisis, 
because that is doing us no good, and simply aims at the most standard 
compatible quirk mode whenever we have a choice. 

Greetings,
Germain