RE: Positioned Layout proposal

[snip]

> 2) It appears easy when it only applies to absolute positioning

Agreed.

> (although
> implementation will not be that easy - it will complicate calculating
> visual bounds, incremental update, pagination and more...).

How does it "complicate" orthogonal algorithms? For example, a visual
bounds test walks the element tree and buildings bounding box from
rectange unions, thus afaics is not concerned with nor aware of (i.e.
orthogonal to) the positioning method is employed.

If pagination algorithm has been designed to be sufficiently orthogonal,
then afaics it should also not be touched for this proposal. Are you
concerned or aware that some layout engines may not be coded with
sufficiently orthogonal (non-spaghetti) design?

I do not know what you mean by "incremental update"? Is that an algorithm
that bypasses full document relayout computation?  Specifics?

> But will it be
> enough? If such a major concept is added to layout model, wouldn't it be
> expected to extend to in-flow content (which of course would be much
> harder, adding circular size dependencies)?

Is that a problem if we restrict that they are only non-float siblings or
parent hierarchy, which are not cols or rows elements? See my prior post
today:

http://lists.w3.org/Archives/Public/www-style/2010Oct/0372.html

Received on Tuesday, 19 October 2010 16:15:24 UTC