Requirements looking good!
Just that I could not find the word norrow in my dictionary and guess you
mean narrow.

A lot of logic is available in XPath-expressions, so specifying that
XPath-expressions be used in certain places would make functions like
count() and substring() available.

In the working-draft we have a lot of elements that are exact copies of
their xsl-counterparts. I think it would be better, both for design and
clarity, if those elements were taken from the xsl-namespace instead. So
that instead of <xupdate:element> we would write <xsl:element>.

What we need for narrowing is a way to select a node-set from the current
node-set. The <xsl:apply-templates select=expression> and the <xsl:for-each>
are ways to do this but the nodes begin to be processed in order
immediately, which is perhaps not what we want (also the name
<xsl:apply-templates> is not good for us).
But then on the other hand, an XPath-expression gives very good control over
which nodes are selected based on properties of the nodes and their
relatives (e.g. children of siblings of ancestors).

Conditional processing is where the processing depends on something other
than the nodes and their relatives. Here <xsl:if> and <xsl:when> would
probably be sufficient.

/Torbj�rn Gannholm
------------------------------------------------------------------
Post a message:          mailto:[EMAIL PROTECTED]
Unsubscribe:             mailto:[EMAIL PROTECTED] adminstrator:    mailto:[EMAIL 
PROTECTED]
Read archived messages:  http://archive.xmldb.org/
------------------------------------------------------------------

Reply via email to