Re: [CSSWG] Minutes and Resolutions 2010-04-21

On Mon, May 24, 2010 at 12:55:17PM -0700, Tab Atkins Jr. wrote:

> Let's see if I can write this up, then:
> 
> "For the purposes of the algorithm, a sibling is the first sibling
> element, pseudo-element, or run of text following or preceding the
> abspos element that is not out-of-flow (in the case of an element) or
> whitespace (in the case of a run of text).  Floats are considered
> in-flow for this purpose."

I suggest s/sibling/immediate sibling/, so as not to conflict with
conform.html#sibling .  (I believe the text already uses "immediate sibling".)


Neither the phrase "in-flow" nor "out-of-flow" is currently defined in CSS2.1;
and "out-of-flow" isn't even used anywhere else in CSS2.1 (unless you count
"[float boxes are] taken out of the flow" or "absolutely positioned boxes are
taken out of the normal flow").  I suspect that all existing uses of "flow" in
CSS2.1 pertain to boxes rather than elements.
Because of all that, readers might either spend time trying to determine what
is meant by "out-of-flow", or might miss cases such as display:none elements.
I suggest either avoiding "in-flow"/"out-of-flow", or giving an actual
definition.

> K, then the rules will look like this:
> 
>  1. If its following sibling is display:table-cell, the auto position
> is the upper left of the border box of the sibling.

I suggest "immediately following sibling" (and similarly for most of the rules)
so as not to conflict with the existing definition of "following sibling" at
that same conform.html#sibling paragraph.


Make sure bidi is handled correctly in use of "left" / "right";
consider using terms "start/end edge" instead (with appropriate definition
provided), which is more likely to be read correctly than having a separate
"swap left & right if rtl" sentence before or after the list.


pjrm.

Received on Wednesday, 26 May 2010 14:20:43 UTC