Input document:

<?xml version='1.0'?>
<article>
<section><para>Paragraph 1</para></section>
<section> <para>Paragraph 2</para></section>
</article>

Stylesheet:

<?xml version='1.0'?>
<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version='1.0'>
<xsl:template match='section/para[1]'>
<sectpara><xsl:apply-templates/></sectpara>
</xsl:template>
<xsl:template match='para'>
<normalpara><xsl:apply-templates/></normalpara>
</xsl:template>
</xsl:stylesheet>

Resin Output:

<?xml version="1.0" encoding="ISO-8859-1"?>
<sectpara>Paragraph 1</sectpara>
<normalpara>Paragraph 2</normalpara>

xsltproc output:

<?xml version="1.0"?>
<sectpara>Paragraph 1</sectpara>
<sectpara>Paragraph 2</sectpara>

xt-xp or xt-xml4j output:

<?xml version="1.0" encoding="utf-8"?>
<sectpara>Paragraph 1</sectpara>
<sectpara>Paragraph 2</sectpara>

As  you see,  Resin  is the  only  XSL processor  which  uses the  wrong
template, just because of the extra space before the second <para> :(

BTW, I did not specify any encoding for the output, whereas Resin uses a
default encoding of ISO-8859-1...  IMHO utf-8 (like xt) or nothing (like
xsltproc) should  be specified as  default.  FWIW, all extra  spaces are
removed from output for readability.

Message cross-posted to  xsl-list for comments on this...  any help will
be greatly appreciated.
-- 
Jean-Baptiste Quenot

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