Re: Tables and CSS (was: With CSS how does one ALIGN block-leve

Neil St.Laurent wrote:

<<Why don't we do as was suggested for centering an entire table and do
vertical align using margin-top and margin-bottom, where center would

margin-top: auto
margin-bottom: auto

Because borders fall within the margin. If vertical margins on cells
adjusted automatically to accommodate content, the top and bottom cell
borders would not align horizontally.

> Frame and border are interrelated. Rules are borders on TH and TD.
> Neither IE nor N fully support these attributes. In IE, cellspacing
> is lost where cells abut without rules..

We don't need a perfect translation though, only one that will
accomodate all combinations, which I think changing the borders on
TD's and TH's would do for FRAME and BORDER from the table.

This seems to be IE's approach. A table is bordered entirely with
combinations of borders on the individual cells. Cellspacing is selectively
ignored to insure that rules are not broken where cells abut.

I would prefer that cellspacing not be ignored, even though this means that
cells with margins that abut without rules would have a gap in rules that
cross the cellspacing, e.g.:

  +---------   -----------+
  |                       |
  | cell a        cell b  |
  |                       |
  +---------   -----------+

Authors wanting continuous rules would use padding, not margins, on cells.
Authors wanting a custom 3D effect could use margins to get it.

There's another complication to this model: in CSS1, vertical margins
collapse, horizontal margins do not. In order to mimic current HTML
cellspacing with margins on cells, horizontal margins on abutting cells
would have to collapse.

<<I was kind of hoping using VOID would create a VOID that sucks the
current browser into it and erases it from the system... no such luck

(I wish there were a void I could throw this version of OE into and have it
come out without some 'convenience' features that actively fight me -- not
sure whether they were designed for Beavis and Butthead or by them.)

<<I think this is still related to the content-model flow of CSS, which
is specifcally top-down.  It appears as though Table cells have an
entirely different flow to them which cannot be accomodated in CSS.>>

Yes, vertical alignment of cells requires intimate interrelations among
elements that are, in the familial sense, first cousins.

David Perrell

Received on Wednesday, 22 October 1997 14:02:50 UTC