> Are you stating that "*[name()=$keyword]" worked for me as 
> the value of a match attribute of xsl:template, using 
> libxslt, quite literally "by accident" (i.e., because the 
> processor doesn't check this constraint)? If so, are there 
> any side effects which may manifest as bugs elsewhere in the 
> stylesheet by taking this terse approach? Do you recommend 
> abandoning this approach?
> 

There are several 1.0 processors that don't enforce this constraint. The
constraint is there in the spec because of a misguided attempt to
prevent circularities such as

<xsl:variable name="x">
  <xsl:apply-templates match="/*/item[1]"/>
</xsl:variable>

<xsl:template match="item[$x]">
  ...
</xsl:template>

Like the restriction on using RTFs as node-sets, it's totally
unnecessary but conformant processors are required to enforce it.

Michael Kay


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