- From: Andrew Fedoniouk <news@terrainformatica.com>
- Date: Mon, 4 Oct 2004 23:40:00 -0700
- To: "www style" <www-style@w3.org>
- Message-ID: <001d01c4aaa6$24122b00$0301a8c0@ATHLON>
There are two attributes which have completely
different meanings in different situations in CSS currently:
vertical-align
1) inline positioning of element itself
2) content alignment in table-cell (and table-row).
direction
1) direction of text flow.
2) order of columns in display:table elements
Sharing name 'vertical-align' between two entities disables
vertical content alignment to be applied to *any* container.
Element can be placed in line and aligned there (e.g. inline-block) and
may have its content aligned vertically at the same time. With
sole 'vertical-align' it is just impossible.
Sharing name 'direction' makes impossible to inherit text direction in table cells.
Proposal:
To introduce
1) 'content-vertical-alignment' attribute with the same set of values as vertical-alignment
2) (probably) 'content-alignment' attribute with left, right, center values. This is not text-alignment
but horizontal alignment of contained blocks.
3) 'flow' attribute with possible values left-to-right, right-to-left, top-to-bottom (default), bottom-to-top(?). This attribute can be applied to *any* block container (e.g. div) . table element accepts only left-to-right and right-to-left values - column ordering.
For the sake of compatibility with HTML/TABLE to introduce also MultiLength units (or %%)
Their formal definition can be found here: http://www.w3.org/TR/html401/types.html#type-length
Availability of such units will be final step allowing to emulate current HTML behavior in CSS in full.
Comment: if we will allow MultiLengths to appear in margins and paddings then
'content-vertical-alignment' and 'content-alignment' can be omitted completely.
(E.g. to align content to bottom it is enough to set container's padding-top:1*;
to top: padding-bottom:1*; and to middle: padding:1* 0; )
If this really make any sense I can formalize this in more strict form for inclusion in the spec.
Andrew Fedoniouk.
http://terrainformatica.com
Received on Tuesday, 5 October 2004 06:40:39 UTC