On Nov 4, 2006, at 08:37, Ian Hickson wrote:
I'm thinking of only allowing integer values, and requiring them to be
equal to the dimensions of the image, if present (and requiring
both to
be present if either is present). Would people be ok with that?
Suppose there are desktop systems in the near future that double the
pixel density of the display compared to what is common today.
Considering the handheld displays Nokia ships, I guess it is only a
matter of time for similarly small pixels to appear in desktop-sized
panels at affordable prices. Apple already seems to be preparing for
this.
To render legacy pages, browsers will probably have to define 1 CCS
px to be equal to two device pixels and render each image pixel of
legacy bitmaps as 2 by 2 device pixels. Now if an author wants to
deliver more precision to such new systems while being compatible
with legacy systems, the reasonable thing to do is leaving the width
and height attributes to values that would make sense for legacy
systems and to quadruple the number of samples in the bitmap by
doubling its pixel height and width.
So I think width and height should not have conformance requirements
tied to pixel dimensions of the references image file. They are
presentational, but they are such a useful presentational
optimization that I think it doesn't make sense to try the get rid of
that presentationalism just to comply with a principle.
--
Henri Sivonen
[EMAIL PROTECTED]
http://hsivonen.iki.fi/
