> I thought that <trigger> of NVDL is good enough for XHTML+XForms
> even when they share the same namespace URI. However, I might
> have been too optimistic.
> Consider an XForms subtree <select1> embededded in an XHTML document.
> In my understanding, another XHTML subtree <p> may be embedded in this
> XForms subtree. (Am I right?)
That's my understanding - a single form can be in multiple parts within a
single HTML document.
> Then, we have to specify "p" using the trigger element of NVDL so that
> we can extract the embedded <p> and validate it against the XHTML
> schema. However, another <p> may appear dicretly below the <body>
> element of the XHTML document. Such <p> will be mistakenly detached
> from the top-level XHTML, although it should not.
> We can overcome this problem by slightly extending the trigger element
> of NVDL. For example, we can allow a <trigger> to specify a list of
> qualified names. This <trigger> element detach an element from its
> parent element when (1) the qualified name of this element is in this
> list and (2) the qualified tag name of the parent element is NOT in
> this list. However, the use of different namespaces for different
> vocabularies is a much better solution.
Can you let us have an example of this new functionality so we can see how
it works in context?