[Home] [By Thread] [By Date] [Recent Entries]


Bullard, Claude L (Len) wrote:
 > It isn't a data island.  It is non-HTML in the renderer.
 > To be specific, it is VML inside HTML using the CSS behavior
 > support and the xmlns declaration, per the manual.
...
 > Then it occurred to me the browser doesn't actually
 > know the difference if it isn't in an XML document.
 >
 > I dunno Joshua.  It seems to me that grandfathering
 > HTML leads to problems worse than actually creating
 > an XML browser that knows it is one.  My particular
 > bug is fixable, but it tells me that when using
 > XML inside non-XML, I can't trust rules I thought
 > I could trust.

IE is actually quite funny when you need to script against the VML object model. If you use elements such as 
v:fill inside a v:shape and set @color on it and then serialize it you will notice that the color attribute 
magically has moved from the v:fill element to the v:shape element (which is the parent of v:fill). It does so 
to a lot of attributes on a lot of child elements and I can't really say I'm particularly fond of it ;).

You could say I have learned the hard way not to trust the IE HTML/"XML" engine at all (although MSXML works 
mostly perfect for my needs in the browser but that is another story).

/Marcus

Site Map | Privacy Policy | Terms of Use | Trademarks
Free Stylus Studio XML Training:
W3C Member