W3C home > Mailing lists > Public > www-style@w3.org > December 1997

RE: Split Horizontal & Vertical Visibility Values

From: Andrew n marshall <amarshal@usc.edu>
Date: Thu, 4 Dec 1997 23:31:56 -0800
Message-ID: <01BD010C.D084C780.amarshal@usc.edu>
To: "'Hakon Lie'" <howcome@w3.org>
Cc: "'www-style@w3.org'" <www-style@w3.org>
On Thursday, December 04, 1997 7:33 AM, Hakon Lie [SMTP:howcome@w3.org] 
> I'm catching up on www-style and the above message seems to be referred
> to several times. I see the need for avoiding reflows, but doesn't the
> 'width' property of CSS1 cover this case? E.g. you can explicitly set
> the width of the element:
>  DIV { width: 10em }
> Further, the proposed 'min-width' and 'max-width' allows the designer
> to set more constraints. Do you have an example which isn't covered
> by these?

Now that you bring up the max-height/max-width concept, I'm having trouble 
determining whether it would be possible.

I'm attempting to created a dynamic sidebar menu tree based upon nested 
<UL>s and associated <LI>s that has a size determined by its content and 
does not have reflow refreshing.  My initial attempts involve toggling the 
display property, but this changes the width of my sidebar if the length of 
the submenu is larger than the compacted tree's width.

What I am now toying with is leaving width to define itself, but 
manipulating height and overflow or height-max. I am having problems 
determining if this is just because IE4 isn't a complete CSS-2 capable 
browser or if the problem remains in the CSS specification.

What I can't do with height and width properties explicitly set a box 
determine it's own size based upon it's content. It was pointed out that 
there are problems with this when it comes to word wrapping, but I 
demonstrated[1] that this has been solved in tables on both major browsers 
acceptably, although I have no concept of how this was implemented. I think 
the concept of an auto formatting is important when designing around the 
differences between monospaced and non-monospaced fonts, or designing a 
multilingual page to deal with phrase lengths in different languages 
(German versus Chinese).

What I have decided is also important is the concept of remaining available 
space as a length unit, similar to the HTML <FRAME> unit '*'. With this, 
the space remaining after such an auto-format could be divided and 
allocated to the surrounding boxes. It also prevent any possible overflow 
from coving adjacent box because they could be size around the auto value. 
This combined with overflow=visible gives a much more precise definition of 
what should happen than any automatic sizing based on content.

[1] http://www.media-electronica.com/~amarshal/Tests/tablebox.html (note 
that no sizing had been applied to the table cells; please ignore the text 
because it was directly copied from an earlier unrelated example)

Andrew n marshall
  student - artist - programmer
      "Everyone a mentor,  Everyone a pupil"
Received on Friday, 5 December 1997 03:12:52 UTC

This archive was generated by hypermail 2.3.1 : Monday, 2 May 2016 14:26:45 UTC