Re: [css3-multicol] column overflow

On Oct 16, 2008, at 3:44 PM, Robert O'Callahan wrote:

> On Fri, Oct 17, 2008 at 8:50 AM, Håkon Wium Lie <howcome@opera.com>  
> wrote:
> Also sprach David Hyatt:
>
>  > Let's say you had way more text in your example, such that you  
> had 3
>  > tiny columns and 24 more columns in the current implementations  
> (all
>  > spilling out horizontally).
>  >
>  > Are you proposing that those 24 extra columns would be stacked
>  > vertically in 8 overflowing rows of 3?
>
> Yes.
>
>  > I think this would be ideal, since I could hit "page down"
>  > scrolling and read each "column page."
>
> Indeed.
>
> It's worth trying, and fairly easy to implement.
>
> One issue is that normally, when you scroll down by a page, Firefox  
> (and I think other apps) scrolls down by a little less than the full  
> page height so that some content that was at the bottom of the  
> previous page is visible at the top of the new page --- this gives  
> users some retained context. So if you do the obvious thing, say  
> html { height:100%; column-width:20em; }, repeated page-down will  
> not put the column tops at a constant offset from the viewport top.  
> We could special-case the page down amount for, say, scrollable  
> elements that use columns and the viewport when the root element is  
> using columns, but that's not a perfect solution.
>
> In general, when you want to put columns on an element that isn't  
> the only child of a scrollable container, I think this is not going  
> to work all that well. I don't know what would, but that's a problem  
> I'd really like to solve.

Yeah that's a good point that there could be other unrelated content  
overflowing (like positioned descendants).

I dislike the idea of adding to overflow-x and overflow-y to supply  
new overflow mechanisms.  To me the values of overflow-x/-y are not  
about what specific type of overflow mechanism should be used, but  
just about whether the overflow is visible, invisible, reachable via  
an automatically appearing/disappearing overflow mechanism, or  
reachable via an overflow mechanism that is always present.

That the overflow mechanism happens to be a scrollbar is really the  
issue, so maybe what's needed is a new CSS property that describes  
more explicitly what the overflow mechanism should be.

dave
(hyatt@apple.com)