OK I went back and read some documentation, and I guess node() really is
supposed to match *all* kinds of nodes.  I suppose it is just that
attributes are not children of their parents (shocking, in a way).  But I
still have to say it seems strange that "@*" (apparently) can match
attributes on the child axis and node() can't ...  I know I am missing all
kinds of subtleties and undoubtedly there were reasons for all of this, and
perhaps the alternatives were even more counter-intuitive, but from a
beginner's perspective, well, there are definitely some quirks to get used
to!

> -----Original Message-----
> From: Mukul Gandhi [mailto:gandhi.mukul@xxxxxxxxx] 
> Sent: Friday, July 03, 2009 9:23 AM
> To: xsl-list@xxxxxxxxxxxxxxxxxxxxxx
> Subject: Re:  returning nodes which have a specific child
> 
> On Fri, Jul 3, 2009 at 6:36 PM, Michael 
> Sokolov<sokolov@xxxxxxxxxxxx> wrote:
> > including the
> > bizarre conceit that attribute *nodes* are not matched by 
> the 'node()' 
> > test (perhaps some other name would have been more appropriate?).
> 
> I am not aware of the discussions, that went amongst the WG 
> members, when XSLT 1.0/XPath 1.0 was designed.
> 
> But I think,
> Calling attributes as "nodes" is ok. As attributes are 
> modeled similar to other kinds of nodes. They have name, 
> value, parent etc.
> 
> Attributes can be referred as @* while other kinds of nodes with
> node() and so on.
> 
> I find this design ok :)
> 
> 
> --
> Regards,
> Mukul Gandhi