[selectors-4] Move quirks mode statements about class and ID selectors to their respective sections. Remove normative dependency on the tree being a DOM tree: that's especially silly given we're giving guidance on mapping other types of trees/structures into the element tree used by Selectors.

Author: fantasai

selectors/Overview.bs

@@ -480,7 +480,7 @@ Selectors Overview</h2>
 		<dt><dfn local-lt="static">static profile</dfn>
 		<dd>
 			The <a>static</a> profile is appropriate for contexts which aren't extremely performance sensitive;
-			in particular, it's appropriate for contexts which evaluate selectors against a static DOM.
+			in particular, it's appropriate for contexts which evaluate selectors against a static document tree.
 			For example, the {{Element/query()}} method defined in [[DOM]] should use the <a>static</a> profile.
 			It includes all of the selectors defined in this document.
 	</dl>
@@ -570,33 +570,19 @@ Structure and Terminology</h3>
 <h3 id='data-model'>
 Data Model</h3>
 
-	Selectors are evaluated against a DOM tree. [[!DOM]]
+	Selectors are evaluated against an element tree such as the DOM. [[!DOM]]
 	Within this specification,
 	this may be referred to as the "document tree".
 
-	In addition to the information present in the document tree,
-	for the purpose of selectors,
-	elements are assumed to have arbitrary <a>pseudo-class</a> flags
-	and <a>pseudo-element</a> slots,
-	as defined by the given <a>pseudo-class</a> and <a>pseudo-element</a>.
-
-	<div class='example'>
-		For example, the first element child of each element
-		has a flag indicating that it matches the '':first-child'' pseudo-class.
-
-		For another example,
-		all elements have ''::before'' pseudo-elements,
-		but not only elements with <a>shadow trees</a> have ''::shadow'' pseudo-elements.
-	</div>
-
-	Five aspects of a DOM element are especially relevant for selectors:
+	Each element may have any of the following five aspects,
+	which can be selected against:
 
 	<ul>
-		<li>The element's type (also known as tagname), which is a string.
-		<li>The element's namespace, which is a string.
-		<li>An ID, which is a string.
-		<li>Classes, which are strings.
-		<li>Attributes, which are pairs of strings consisting of an attribute name and an attribute value.
+		<li>The element's type (also known as its tag name).
+		<li>The element's namespace.
+		<li>An ID.
+		<li>Classes (named groups) to which it belongs.
+		<li>Attributes, which are name-value pairs.
 	</ul>
 
 	While individual elements may lack any of the above features,
@@ -611,10 +597,22 @@ Data Model</h3>
 		and can't be matched by <em>any</em> <a>pseudo-class</a> except for '':host'' and '':host-context()''.)
 	</div>
 
-	When matching a selector against a document which is in <a>quirks mode</a>,
-	class and ID selectors must be matched <a lt="ASCII case-insensitive">ASCII case-insensitively</a>
-	against the classes and ID of the elements in the document.
+	In addition to the information present in the document tree,
+	for the purpose of selectors,
+	elements are assumed to have arbitrary <a>pseudo-class</a> flags
+	and <a>pseudo-element</a> slots,
+	as defined by the given <a>pseudo-class</a> and <a>pseudo-element</a>.
+
+	<div class='example'>
+		For example, the first element child of each element
+		has a flag indicating that it matches the '':first-child'' pseudo-class.
+
+		For another example,
+		all elements have ''::before'' pseudo-elements,
+		but not only elements with <a>shadow trees</a> have ''::shadow'' pseudo-elements.
+	</div>
 
+	Issue: This section is imho unnecessary, and asserting the existence of a "first-child" flag belonging to the element is a) silly, because that information is already in the tree and b) confusing, because one could then interpret that as meaning the element keeps its first-child flag even when it is moved in the DOM.
 
 <h3 id="scoping">
 Scoped Selectors</h3>
@@ -1614,6 +1612,10 @@ Class selectors</h3>
 	be reached, however, so this behavior is explicitly non-normative in
 	this specification.
 
+	When matching against a document which is in <a>quirks mode</a>,
+	class names must be matched <a lt="ASCII case-insensitive">ASCII case-insensitively</a>;
+	class selectors are otherwise case-sensitive.
+
 <h3 id="id-selectors">
 ID selectors</h3>
 
@@ -1667,6 +1669,10 @@ ID selectors</h3>
 	selector. Such a situation could be reached using mixtures of xml:id,
 	DOM3 Core, XML DTDs, and namespace-specific knowledge.
 
+	When matching against a document which is in <a>quirks mode</a>,
+	IDs must be matched <a lt="ASCII case-insensitive">ASCII case-insensitively</a>;
+	ID selectors are otherwise case-sensitive.
+
 <h2 id="linguistic-pseudos">
 Linguistic Pseudo-classes</h2>
 
@@ -2334,7 +2340,7 @@ Tree-Structural pseudo-classes</h2>
 	of children of its parent. When calculating the position of an element in
 	the list of children of its parent, the index numbering starts at 1.
 
-	The <a>structural pseudo-classes</a> only apply to elements in the DOM;
+	The <a>structural pseudo-classes</a> only apply to elements in the document tree;
 	they must never match <a>pseudo-elements</a>.
 
 
@@ -3354,9 +3360,11 @@ Match a Selector Against an Element</h3>
 	</ol>
 
 <h2 id='dom-mapping'>
-Appendix A: Guidance on Mapping Document Languages to DOM</h2>
+Appendix A: Guidance on Mapping Source Documents &amp; Data to an Element Tree</h2>
+
+	<em>This section is informative.</em>
 
-	The DOM document structure is powerful and useful,
+	The element tree structure described by the DOM is powerful and useful,
 	but generic enough to model pretty much any langauge that describes tree-based data
 	(or even graph-based, with a suitable interpretation).