I have an XSLT script collection that converts messed up XML to usuable XML. 
Instead of generic template matching, I want to establish a context node for 
some branches so that I don't have to rely on absolute path. Here's a 
simplified version of my work:


[root.xsl]:
    <xsl:include href="common.xsl"/>

    <xsl:template match="/">
        <patientDisk>
            <xsl:call-template name="patientData"/>
        </patientDisk>
    </xsl:template>

==================================

[patientData.xsl]:

    <xsl:template name="patientData">
        <birthyear>
            <xsl:value-of 
select="document('rpt_pat.xml')//parameter[@name='PatientBirthYear']"/>
        </birthyear>
        <gender>
            <xsl:value-of 
select="document('rpt_pat.xml')//parameter[@name='PatientGender]"/>
        </gender>
    </xsl:template>


Instead of specifying document('rpt_pat.xml') every time using absolute 
paths, I could have used <xsl:for-each select=document('rpt_pat.xml')>. 
However, I'd like to avoid using this syntax because there's only one 
instance of document('rpt_pat.xml') node. Using <xsl:for-each> can lead to 
whoever that reads my XSLT scripts into thinking that there are multiple 
instances of the document('rpt_pat.xml') node, adding confusion to the 
already huge chaos.

I'd like to avoid <xsl:template match="(something)"> because some tags have 
the same names in the messy XML files that I'm trying to parse, and they can 
be distinguished only by observing the hieracy (which branch that XML tag is 
under).

Can anybody suggest a neat way to do that? Thanks in advance.

-- Hoi