Re: top-level element enigma

Ian Hickson wrote:

>Well, isn't this what the @page {} rule does? (CSS2, page.html#didx-page)
>I suppose one could also have an @screen or an @canvas.

@page, as described, doesn't work for accessing the canvas. It doesn't
explain reverse-inheritance of background, nor does it allow setting the
default font for display. And it doesn't explain scrolling of the
background. @canvas, though, makes sense to me.

The canvas is described, along with the use of the HTML element for setting
backgrounds, at <http://www.w3.org/TR/WD-CSS2/media.html#h-8.3.1>, but this
description is somewhat confusing in that it unnecessarily introduces new
and sometimes conflicting terminology. The first paragraph under 'Scrollable
media' is just plain wrong. There should be no instance where the canvas
does not fill the viewport. If the canvas doesn't fill the viewport, what
background is used for the empty areas?

Put more simply, the canvas has fixed height and width corresponding to the
UA window or frame (viewport) and an auto scroll property. It is the canvas
that scrolls, not the BODY element, and by default it may scroll in both
directions. There should be no instance where the canvas does not fill the
viewport. If the canvas doesn't fill the viewport, what is the background of
empty areas?

Whether @canvas or HTML, there needs to be a parent of BODY that fits the
above description. That way, BODY can be absolutely or relative positioned
and sized such that the markup that produces the effect on
<http://www.hpaa.com/css1/div_frame.html> when using IE4 as UA will be
condoned by the CSS spec (I believe that relative sizes as applied to BODY
are currently undefined.)

David Perrell

Received on Friday, 19 December 1997 17:05:46 UTC