Re: Float overflowing behavior

On Thursday, August 15, 2002, 9:44:09 AM, you wrote:
> it would be nice to suggest via CSS that containers are supposed to
> fully enclose their contained elements, even if a contained element
> has been suggested to 'float' ... IMO though; it seems to be more
> logical if such a property is applied to the containing element
> instead of the contained element.

I'm split on this one.

A hypothetical float-overflow property makes a lot of logical sense to
me when applied to a containing element (similar to overflow), but
would that then effect *all* floats that are descendants of that element?

If so, would that make something like this (excuse the nasty ASCII
art, I hope nobody is using a screenreader on this list)...

   +---------+  +---------------------------+ 
   | float 1 |  | float 2 start...          | 
   +---------+  |                           | 
  +==============|                           |=+
                 |                           |
                 |            ...float 2 end |

...impossible to achieve using floats?

If the property was applied to the floats individually, you could set
one to overflow its container and one to not (I believe the above effect is
currently impossible to do using floats [without knowing the set
height of float 1], even when using the current overflow hack) - it
does complicate things a bit, however.

It seems a lot of people have ideas for extra things to add to the
float specification, and everyone seems to have their own view on what
floats should be for.

Some view them as simply a handy way to move images to the side of
text, others (including me) view them more as a full layout tool - and
a good replacement for tables. I'm almost wondering if they could do
with a separate CSS3 module (or some kind of sub-module)...

Although, I have been in London all day and am now very, very tired -
so the above may well be a meaningless ramble - sorry if so :)

Tom Gilder

Received on Thursday, 15 August 2002 18:59:02 UTC