W3C home > Mailing lists > Public > www-style@w3.org > July 2003

Re: interpretation of float model - Mozilla vs Opera & IE

From: L. David Baron <dbaron@dbaron.org>
Date: Mon, 21 Jul 2003 10:33:58 -0700
To: www-style@w3.org
Message-ID: <20030721173358.GA21308@pool-151-197-214-85.phil.east.verizon.net>

On Monday 2003-07-21 18:00 +0100, Afternoon wrote:
> I firmly agree with this. It goes back to the discussion last week.  
> There is no other way to achieve this effect in CSS except through the  
> table syntax. The effect is very useful.

It seems to me that if you want that effect, the concept you're
describing is tabular layout, not floats.  The use of floats to create
things that look like tabular layout is fragile, and would become
considerably more fragile if this proposal were accepted.   After all,
floats weren't designed for this type of thing -- they were designed for
the idea of taking something out of the flow so that text would wrap
around it.

To see why layouts using floats would become more fragile, consider what
happens when you resize the window while looking at the following
testcase:
  http://www.hixie.ch/tests/adhoc/css/box/float/008.html
in IE for Windows, to a width just larger than the width that allows one
float to fit on the page without wrapping.  Windows IE puts a second
float next to the first, and makes it very tall.  This would mean
authors have a great way to create pages that look like they work fine,
but in reality break very badly at certain window widths.  This is a bad
thing.

If you want to be able to do tabular layouts using CSS, it would make
much more sense to encourage browser makers to support the table display
types than to make floats even stranger than they already are.

-David

-- 
L. David Baron                                <URL: http://dbaron.org/ >
Received on Monday, 21 July 2003 13:36:42 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:54:22 GMT