But speaking to the larger question; even if most popular browsers (or
browser-like OS elements) will display the properties of the :hover
pseudoclass on any element, would it be wise to apply them? It seems
to me that users are becoming accustomed to associating the :hover
properties with hyperlinks. Just as it would be valid, but unwise, to
make ordinary text appear blue and underlined (even if the author
might have good reason for doing so) I suggest it would also be
unwise to apply :hover properties to anything except links.
This is good advice in itself, indeed, but I think it missed the point.
Or at least it did as far as my bit was concerned.
Here I inherited a page that went something like this - in the scope of a
class let's say "mymenu":
<td><div><a href="...">text</a></div></td>
The desire was to specify a :hover behaviour *for the link* which involved
also some CSS properties that are only valid for block-level elements.
The associated CSS wanted, really, to specify something like
.mymenu div:hover { "border-color change" }
and this would've worked on Mozzi etc. - but not on MessIE.
So the author of this lump of code had instead used an "onmouseover/
onmouseout", calling out some javascript in the HTML. For every
goddamned cell in this column of the menu. It was just too much.
That's why I was interested to know (haven't managed to get it to work
myself yet) whether MessIE could be persuaded to do this on the a:hover
itself - if only one could define the "a" as a block-level element within
its TD.
I'll be honest - myself I think this hover stuff is just a fad, that next
year will look as dated as HTML/3.2. But the sponsor of this site is very
happy that the links are almost indistinguishable from the surrounding
text on a typical graphical browser, until one waves a mouse at them,
whereupon they light up like Blackpool Tower - won't accept any change
in its visual behaviour. Oh well - don't mind me, I'm getting old and
crabby. I'm sure the new students at whom these pages are aimed will be
impressed.