> > At one end of the spectrum, by making minimal use of padding and
> > (ie setting their dimension to zero) and not relying on
> > alignment between block-level elements, you can usually ignore
> > cross-browser box-model differences entirely.
> Is this acheivable. My golden rule idea is a vanilla flavor, that
> cross browser. I am guessing that this first approach uses float and
> absoulte positioning. Does this limit the complexity of the design
> template. IE can this approach permit vanilla designs with multiple
> nested divs etc.
If you set border:0px and padding:0px, then the width and height issues
go away. So a vanilla rule to always use zero border and padding makes
sense if you are trying to eliminate cross-browser issues... it's just
awfully confining to almost any design.
> > At the other end, SNIP
> >From this I deduce that you feel that padding is not a postioning
> a visual one. you don't mention margins. I guess I get confused
> can move a box around by changing any of the parameters. border (width
> height) line-height, margin, padding, float and absolute positioning.
> If what you are saying holds true some of these should be reserved for
> visual presentation and others for positioning.
> so knowing this difference could be a rule.
> (the interaction of many things effect position, but if I remove from
> bag of tricks things that are not ment to be for positioning maybe I
> gain some consistancy???)
I don't think you can peel this apart. Padding and border, when
present, effect positioning, and can be used for visual if you give
them color. IE prior to 6.0/strict incorrectly includes them inside
"width"... correct implementation is that they are outside width. So
for IE <6, total width=width+margin(s). For compliant browsers, total
width=width+padding(s)+border(s)+margin(s). Ditto for height.
> > So.... my SNIP
> Ok good, but is there a way to make a vanilla that side steps the
> browser issues.
> It could be the answer is no. But at least it will be possible to pin
> factors down a bit.