Re: [selectors4] Proposal for extending Selectors Level 4 with possibility to select attributes

On 11/11/13 2:54 PM, Domenic Denicola wrote:
> I think the issue in question here is that in HTML and in the DOM used by browsers, an attribute is not a node.

Actually, right now it is.  A broken and incorrectly implemented and 
generally useless node, typically, but a node nevertheless.  So far.

The reasons for making it not be a node largely have to do with not 
wanting to have broken and incorrectly implemented and generally useless 
stuff in the plaform.

> Whereas of course in XML and in the XML DOM, it is.

There is no separate "XML DOM".  There is just the DOM; it's the same 
for HTML and XML in practice.

> If you wish to use HTML APIs like querySelector or querySelectorAll

Those aren't HTML APIs.

> I believe the abstract you are referencing uses the HTML definition of node, thus the confusion.

The abstract being referenced is informative text, and hence no one 
really cared all that much about the exact wording of it.  I wouldn't 
read anything in particular into its wording, other than the usual "The 
CSS WG couldn't quite decide how much it wants to depend on the DOM and 
how much it wants to try to define things in terms of abstract trees of 
which the only implementation just happens to be the DOM".

-Boris

Received on Monday, 11 November 2013 20:11:27 UTC