Re: layout idea

On Mar 20, 2009, at 2:24 PM, Brad Kemper wrote:

>
>
> On Thu, Mar 19, 2009 at 12:34 PM, David Hyatt <hyatt@apple.com> wrote:
>
> I think the model should encourage not specifying display:table-cell  
> cells at all.... i.e., objects that specify table-position should  
> commonly just be blocks not cells.
>
> Sorry to have to disagree on this point too. It is likely that I  
> will have an ID or CLASS on the element that I want to be a header  
> or sidebar or whatever, and then I can use that to not only make it  
> into a table-cell, but to style it like one too. This means that I  
> can have it's background extend to the full height of the table, for  
> instance, or have its border shared with another cell and centered  
> along that edge, while taking advantage of the sizing qualities of  
> that non-fixed tables allow. All in one rule with that ID selector.  
> I would almost always want the DIVision I select to become the cell,  
> with background, border, shadows and all applying to the whole cell,  
> and not just be stuck in as a block at the top of the cell.

I said "encourage" not "disallow."  If you look at my other comments,  
I'm in favor of being able to specify table-position on display:table- 
cell objects as well as non-table-cell objects.  It sounds like you  
would rather there just never be any anonymous cells in these grids.

If table-position only works on table cells, then you end up being  
unable to control where non-table-cell objects end up, since you have  
no control over where the anonymous cell created to wrap those objects  
gets placed in the grid.  You end up with source-order dependence  
again unless you're willing to explicitly specify every single table  
cell in your source.  That might be ok, but I would need to think  
about it some more.  My gut feeling is that there are situations where  
you would want objects to share a cell while still being able to  
maintain source order independence.

dave
(hyatt@apple.com)