In my opinion, the influence of the white space nodes on the content
elements should be made clear in the spec. Consider the following two
XML samples where p is a bound element.
<p><a/></p>
and
<p>
<a/>
</p>
The explicit children of the p elements are different. The first <p> has
only <a>, but the second one has <textnode>, <a>, and <textnode>. If the
<a> element is included into the shadow tree through includes attribute,
the next content element will be replaced either with the white spaces
or with its own content depending on the used markup.
That may confuse authors since they usually use the white spaces to
clarify the markup. I think it should be mentioned that the white spaces
in the bound element's markup may affect to the shadow tree.
Mikko
