W3C home > Mailing lists > Public > www-style@w3.org > September 2011

RE: [css3-flexbox] Too many flex-flow values?

From: Alex Mogilevsky <alexmog@microsoft.com>
Date: Fri, 9 Sep 2011 21:29:13 +0000
To: David Hyatt <hyatt@apple.com>, Tab Atkins Jr. <jackalmage@gmail.com>
CC: www-style list <www-style@w3.org>, Phil Cupp <pcupp@microsoft.com>
Message-ID: <D51C9E849DDD0D4EA38C2E539856928412124F0C@TK5EX14MBXC213.redmond.corp.microsoft.com>
I agree, perhaps not as strongly, but it will be fine without physical flow. Grid doesn't have physical orientation, grid in WPF and Silverlight doesn't have it either.

There is another option though - remove logical. As much as we like the world to move towards logical, the rest of CSS is not ready. Even though we can say

	.box { display:flexbox; flex-flow:column; }

We can't say

	.box > * { margin-after:0.5em; }

So if design wants to adapt to writing mode change it will most likely have to have alternative styles for vertical anyway.

Somehow I have a feeling that proposal to remove logical won't fly, just want to share the thought that any local logical/physical solution will eventually have to be revisited (when there is finally a way to define margin-after or margin-inside).


-----Original Message-----
From: David Hyatt [mailto:hyatt@apple.com] 
Sent: Friday, September 09, 2011 1:51 PM
To: Tab Atkins Jr.
Cc: Alex Mogilevsky; www-style list
Subject: Re: [css3-flexbox] Too many flex-flow values?

I strongly support the idea of just dropping the physical values. I don't think they are necessary.


On Sep 8, 2011, at 10:56 AM, Tab Atkins Jr. wrote:

> On Wed, Sep 7, 2011 at 7:26 PM, Alex Mogilevsky <alexmog@microsoft.com> wrote:
>> Good question. Technically you are right, any style that uses "horizontal" or "vertical" can have an override in style sheets for different writing modes, and it should be extremely rare -- I haven't seen many designs that easily transform with direction change.
>> However... the only thing this change would simplify is the chapter you are writing. It won't be more intuitive to say "row" or "column" for designs that are not tolerant to 90 degree rotation (vast majority of them), it will be a hassle for designs that are (think about a combo box control), and it makes no measurable difference for implementation complexity.
> Do we think the same is true of Grid?  I'd like to stay consistent 
> with that spec.  Right now, if you lay out your page with Grid, the 
> entire layout gets mirrored in rtl, and rotated in ttb.  If the vast 
> majority of layouts don't make sense to be rotated, Grid should either 
> have a similar control (and we should match syntax), or it should have 
> a single physical-based orientation (presumably with cell (1,1) being 
> in the top-right).
> Alternately, if we think that the vast majority of designs aren't 
> tolerant to rotation, then we should probably remove the writing-mode 
> dependent values entirely, and change the default to 'horizontal'.
> (On further thought, the issue I put in the spec about dropping the 
> mixed values is wrong - they're the values that depend on 'direction', 
> and you've demonstrated use-cases for them.)
> ~TJ
Received on Friday, 9 September 2011 21:30:21 UTC

This archive was generated by hypermail 2.4.0 : Monday, 23 January 2023 02:14:03 UTC