Network diagram - node set intersection

==========> Direction of graph
        ______
            |        |
          / |__B__|
______/           \_____
|        |                |        |
|__A__|           |__D__|
          \                /
           \ ______/
           |        |
           |__C__|

I'm working on producing a network diagram in SVG by processing an XML document 
with XSLT. Above is a symbolic representation of the output. The relationship 
between nodes in the graph can't be represented directly in XML because it 
isn't a tree (Node D has two parents which isn't allowed in XML).

Here is a simplified sample of the XML document:
<network>
  <node>
    <node-id>A</node-id>
    <predecessor-id></predecessor-id>
    <successor-id>B</successor-id>
  </node>
  <node>
    <node-id>B</node-id>
    <predecessor-id>A</predecessor-id>
    <successor-id>D</successor-id>
  </node>
  <node>
    <node-id>C</node-id>
    <predecessor-id>A</predecessor-id>
    <successor-id>D</successor-id>
  </node>
  <node>
    <node-id>D</node-id>
    <predecessor-id>B</predecessor-id>
    <predecessor-id>C</predecessor-id>
    <successor-id></successor-id>
  </node>
</network>

In my stylesheet I have a template that matches <node>. For the purpose of 
positioning the SVG elements I want to determine how many other nodes have a 
<predecessor-id> that matches one of the <predecessor-id> children of the 
context <node>. In addition to determining how many "siblings" a <node> has, I 
also want to know where the context node is in document order relation to these 
"siblings".

The <node> elements with common <predecessor-id> values would be, in the 
context of the network diagram, "siblings". Of course in the XML document all 
the <node> elements are siblings in the XPath sense.

To determine the number of siblings a <node> element has, I thought I would 
count the elements in the intersection of the set of <predecessor-id> children 
of the context node with the set of <predecessor-id> children of all the other 
<node> elements. I modelled the expression on the example shown on page 425 of 
the XSLT Programmer's Reference 2nd Edition.

Here is a simplfied stylesheet:

<?xml version="1.0" encoding="UTF-8" ?>
<xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform";>
  <xsl:output method="xml" indent="yes" encoding="UTF-8" />

  <xsl:template match="/">
    <xsl:apply-templates />
  </xsl:template>

  <xsl:template match="network">
    <xsl:apply-templates />
  </xsl:template>

  <xsl:template match="node">
    <xsl:variable name="this-node-id" select="node-id" />
    <xsl:variable name="this-predecessor-nodes" select="predecessor-id" />
    <xsl:variable name="other-predecessor-nodes" select="/network/node[node-id 
!= $this-node-id]" />
    <xsl:variable name="sibling-cnt" 
select="count($this-predecessor-nodes[count(. | $other-predecessor-nodes) != 
count($other-predecessor-nodes)])" />

    This node id = <xsl:value-of select="$this-node-id" />
    Sibling count = <xsl:value-of select="$sibling-cnt" />
  </xsl:template>

</xsl:stylesheet>

This yielded the following output:

<?xml version="1.0" encoding="UTF-8"?>
  

    This node id = A
    Sibling count = 1
  

    This node id = B
    Sibling count = 1
  

    This node id = C
    Sibling count = 1
  

    This node id = D
    Sibling count = 2

It the intersection operation didn't do what I expected. I expected that nodes 
A and D would have a Sibling count of 0 and that nodes B and C would each have 
a sibling count of 1.

Could anyone tell me where I'm going wrong and point me in the right direction?
Thanks.

-- 
Charles Knell
cknell(_at_)onebox(_dot_)com - email

 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list