Re: The future of XML related technologies in Mozilla

Ian Hickson wrote:
> On Wed, 1 Nov 2006, Boris Zbarsky wrote:
>> OK, I sense a significant communications failure here.  Let's say we 
>> have:
>>
>>   <div id="foo">Text</div>
>>
>> and script as follows:
>>
>>   var node = document.getElementById("foo");
>>   var str = (new DOMSerializer()).serializeToString(node);
>>   var str2 = node.innerHTML.
>>
>> Then str is the string:  '<div id="foo">Text</div>' while str2 is the 
>> string:  'Text'
>>
>> To get the string in |str| using innerHTML, you would have to use the 
>> code with importNode and company that I posted in my previous post. This 
>> is why outerHTML exists, fwiw.
> 
> Oh, I see what you're saying. I have no objection to introducing outerHTML 
> as well. Currently I haven't done so mostly because Mozilla doesn't 
> support it, so I assumed there wasn't strong demand for it.

The problem with outerHTML is that it makes sense to get, but is very 
confusing to set. I.e setting outerHTML removes the current node from 
the document so any additional operations on the node after that are 
effectively no-ops. I.e. something like

node = document.getElementById("foo");
node.outerHTML = "<ol><li>hi</li></ol>";
node.setAttribute("start", 5);

won't do what many people would think it does. Of course, we could make 
outerHTML readonly, but that'd be confusing because innerHTML is not, 
and outerHTML is not in IE.

/ Jonas