Re: Another view (sorry) on XBL and behaviours

What Ian and David are proposing is such a complex merging of layers, that
they themselves apparently do not see all the bad things that can happen.
One example of many follows...

>David Hyatt wrote:
>All other solutions that I've heard involve polluting the main
>document with these built-in event handlers and behaviors. 
>For example, how would you handle easily defining multiple 
>complex presentations and switching between them without 
>leveraging alternate stylesheets?


>Ian Hickson wrote:
>   code { binding: url(code.xml#brief); } 
>   code:hover { binding: url(code.xml#details); } 


I was waiting to see if any one else would point out that Ian proposed a
logical endless loop condition.

When you dynamically replace the implementation (binding), then the
bounding box can change, then the :hover state can osscillate.  You could
send the UA into an endless loop.

This is just one of many examples, wherein when you violate the separation
of presentation and markup, then all kinds of unpredictable, _CATASTROPHIC_
things will happen.  Please, I urge everyone to read "Shelby's Final
Position Paper on XBL"[1].  The core design issue is whether you want to
combine things that by design were separated a few years ago with the birth
of CSS.

BTW, apologies for breaking my promise not to post again on this thread,
but I did not see any one else point out that which was obvious to me and
also very important.

I strongly advise the people who are coding in XBL now, to consider the
future of their work.  They could switch to XSLT now.

(Again I am not against XUL conceptually.  Programming new "standard" tags
into the UA is a valid activity that has been happening since beginning of
HTML).

I intend this post as no disrespect to David Hyatt and thank him for his
well focused points.

-Shelby Moore

Received on Tuesday, 7 January 2003 05:00:54 UTC