W3C home > Mailing lists > Public > www-ql@w3.org > April to June 2001

Re: What's implementable in XPath expressions?

From: Michael Dyck <MichaelDyck@home.com>
Date: Fri, 27 Apr 2001 00:14:01 -0700
Message-ID: <3AE91C39.C02E2FD6@home.com>
To: www-ql@w3.org
Howard Katz wrote:
> 
> I've just been having some conceptual difficulty grokking the
> marriage of an angle-bracket representation of XML with XPath
> relative location syntax. Unaccustomed to the mixed notation, I've been
> finding the idea of "/<aRoot></aRoot>/someRelPath" difficult to grasp.

Ah, well, *that* example is another matter. According to the Appendix B
grammar, the only thing that can come after a '/' is a Step, which is fairly
restricted and does not include ElementConstructor. So 
    /<aRoot></aRoot>
in your example is syntactically invalid. However, an ElementContructor can
occur *before* the first '/' of a PathExpr, so
    <aRoot></aRoot>/someRelPath
is syntactically okay. (Although because the constructed aRoot element has
neither children, attributes, or parent, there's not much you could
substitute for "someRelPath" that would actually yield anything. Perhaps
just dot.)

> Once I translate that into "/aRoot/someRelPath," I'm fine.

Nope, not the same thing. Or rather, if we transpose to the land of
syntactically valid expressions,
    <aRoot></aRoot>/someRelPath
is not equivalent to
    aRoot/someRelPath

-Michael
Received on Friday, 27 April 2001 03:15:48 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 22 July 2006 00:10:17 GMT