On Sat, 4 Mar 2006, Siggi wrote:
"Spartanicus" <in*****@invalid.invalid> schrieb im Newsbeitrag
a) the image Sample.jpg in its original size, irrespective of
values for "width" and "height"
Only in IE, IE doesn't really support image embedding via the
object element.
Thanks! With Netscape no problem: no scrolbars, and image sizeable!
Despite what had seemed to me to be "obvious" theoretical advantages
of <object> over <img>, I concluded some years back that the browser
developers (at that time it was between Netscape 4 and IE-something)
had so royally screwed-up the implementation of <object> as to make it
unusable in practice...
I'm afraid I hadn't re-examined the situation since; however, <object>
provides a fallback, intended for browsers which don't support it, and
that fallback could be an <img>. And the <img> in its turn can
provide a text fallback.
<object ...>
<img ... alt="alternative text">
</object>
OK, so that isn't ideal as a way of providing a text fallback, since
one of the motives of <object> was to permit a fully marked-up piece
of textual content as its fallback, whereas here we've had to accept
one of the very limitations for which <img> was supposed to be
phased-out, namely the fact that its textual alternative has to be an
*attribute value*, not a piece of marked-up content. But let's follow
the idea a little way, nevertheless.
Maybe there's a way of coding <object>, with <img> as its fallback, in
such a way that web-compatible browsers behave correctly, while MSIE
renders the <img> instead?
If it's only MSIE that needs to be shielded from the <object>
nowadays, then surely MSIE's "conditional comments" can be used, to
hide the object markup from MSIE and leave it seeing only the <img> ?
A web search suggests
http://diveintomark.org/archives/200...ernet_explorer
(but as you see, this was from 2003).
It also suggests
http://www.w3.org/TR/2004/WD-WCAG20-...ortdescription
and the immediately following items, which are rather depressing:
"User agent problems make this technique ineffective at this time."
And, on the other hand, unless one is using some kind of custom
processor for generating the (x)HTML, there's the question whether
this is really worth all the effort, seeing that (even as WCAG2
admits), rather few of the potential benefits of <object> can be
realised in a compatible fashion? If you want zoomable images, most
of the graphical browsers browsers I tried nowadays support <img>
sized via CSS in em units, even if their zooming algorithms don't
produce top quality results.
Just some thoughts, anyway. Further study needed?
