W3C home > Mailing lists > Public > www-style@w3.org > June 2004

Re: CSS3 and better support for table free designs

From: Dave Hyatt <hyatt@apple.com>
Date: Sun, 27 Jun 2004 02:22:00 -0700
Message-Id: <71CF70B8-C81B-11D8-B7A6-000A95AE8B18@apple.com>
Cc: www-style@w3.org
To: Andrew Fedoniouk <news@terrainformatica.com>

I agree CSS3 needs a spring and strut layout system, and I'm working on 
a module that describes the XUL box layout system (implemented in both 
Mozilla and Safari).  I'll reiterate one more time that I believe 
overloading width has a fundamental flaw, since you effectively prevent 
flexible layouts from laying out objects at their intrinsic widths (or 
specified widths) before the flex is applied.  This is critical for 
example when opening dialogs to an intrinsic size, but still allowing 
them to resize.  The flexing doesn't kick in until the user resizes the 
window, and the system still needs to have a notion of preferred 
intrinsic width to use for initial display.

As for flex applied to the other components of the box, I think that's 
better done with new CSS properties that can insert flexing springs 
before and after the margin box.  I don't think margins should be 
overloaded.  Flex is not a unit of length, and should not be 
represented as such in CSS.

I also think it's better to just use a new display type in CSS and not 
to attempt to overload normal block and line layout with this behavior. 
  It's far easier to describe and far less complex if this feature is 
done with new display types in a separate CSS3 module.

dave

On Jun 26, 2004, at 4:35 PM, Andrew Fedoniouk wrote:

>
>> ... Andrew keeps pushing this proposal of his again and again and 
>> again
> ...
>
> :) 'cause it is simple and natural.
>
> Current state of CSS is the "Procrustean bed" for modern online
> applications:
> it perfectly suits paradigm "html-as-an-endless-tape" - static texts,
> documentation, etc - one dimension layout pretty well.
> But it has merely nothing for support two dimensional flexible layouts 
> - the
> must for online applications having forms, interactive components and 
> other
> stuff bound with viewport rectangle.
> "Positioning" of different kinds is an attempt to provide a "solution" 
> for
> such layouts, IMO, failed as it (positioning) practically cannot be 
> used
> without scripting (that is another song).
>
> %% units or XUL's "flex"es can help a lot to handle such layouts.
> I like %% length units more (objectively) as they allow make flexible 
> not
> only width/height (as it is in XUL now) but also  other dimensional
> attributes. E.g. in one of my examples I am using:
>
> <div style="width:XXpx; height:YYpx">
>   <div>Text at top of a container</div>
>   <div style="margin-top:100%%" >Text at bottom of a container</div>
> </div>
> to place two lines at top and bottom of a container.
>
> Easy and natural (from human and CSS point of view), isn't it?
> Just try to make something like this in current CSS. No way.
>
> Other great benefit from %% units - they will finally remove misuse of
> text-align and vertical-align.
>
> See, for table-cell, vertical-align means *alignment of container 
> content*
> and for IMG vertical-align means exactly
> opposite - *alignment of the particular element in its container*. Not 
> bad,
> huh?
>
> See:
> <div style="padding-top:100%%">...</div> - content of the div will be 
> bottom
> aligned
> <div style="padding-bottom:100%%">...</div> - content of the div will 
> be top
> aligned
> <div style="padding:50%% xx">...</div> - content of the div will be in 
> its
> middle.
>
> Sorry guys, I don't really want to agitate your more....
> For me benefits are clear, mechanisms are proven to work, example was
> published - anyone can try by himeself.
> Yes, my English is not so strong to write a formal explanation. It 
> would be
> just great if somebody has spare time and can help me to do this. I'll
> appreciate a lot.
>
> Andrew Fedoniouk.
> http://terrainformatica.com
>
>
> ----- Original Message -----
> From: "Christoph Päper" <christoph.paeper@tu-clausthal.de>
> To: <www-style@w3.org>
> Sent: Saturday, June 26, 2004 12:57 PM
> Subject: Re: CSS3 and better support for table free designs
>
>
>>
>> <ooar123@ntlworld.com>:
>>> On Fri, 25 Jun 2004 21:47:35 -0700, "Andrew Fedoniouk"
>>>
>>>> body { flow: horizontal; height:100%; margin:0 }
>>>> #left { width: 100px; }
>>>> #middle { width: 100%% } - will take rest left from #left and #right
>>>> #right  { width: 200px }
>>>
>>> I tested the above exactly and it didn't work-
>>
>> Of course not, because ...
>>
>>>> 'flow' and  %% units are my personal invention - they are non 
>>>> standard.
>>
>> ... Andrew keeps pushing this proposal of his again and again and 
>> again
> ...
>>
>> The CSS2 solution is:
>>
>>    html     {height: 100%; display: table;}
>>    body                   {display: table-row;)
>>    #left, #middle, #right {display: table-cell;}
>>    #left    {width: 100px;}
>>    #right   {width: 200px;}
>>
>> This fulfills all your requirements except:
>>
>> | serve the columns in what ever order I wish.
>>
>
Received on Sunday, 27 June 2004 05:22:32 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:54:30 GMT