RE: Paper on symmetry of XPath axes

Many thanks for forwarding this paper, I think it's an excellent
contribution to the state of the art in XPath optimization (and also a good
improvement on the draft that you sent me privately).

Am I right in thinking that you don't address the problem of positional
predicates on a reverse axis, the typical example being
preceding-sibling::h2[1]? I suspect these can be tackled by an extended
version of your algorithm, but it seems to me they aren't handled correctly
by the rewrite rules you currently define.

I'm also uncertain about the extent of applicability of these rewrite rules.
Certainly one can envisage a scenario in which it's valuable to be able to
evaluate a single XPath expression in a single forwards scan of a document.
I'm not sure how well this extrapolates to a scenario like XSLT, where there
are many XPath expressions, each one starting at nodes selected by previous
XPath evaluations.

Neverthless, this is a very valuable contribution to the repertoire of
techniques available, so thanks for making it available.

Mike Kay

> -----Original Message-----
> From: Holger Meuss [mailto:meuss@cis.uni-muenchen.de]
> Sent: 11 February 2002 15:46
> To: www-xpath-comments@w3.org
> Subject: Paper on symmetry of XPath axes
> 
> 
> Dear colleagues,
> 
> my colleagues and I investigated the issue of symmetry in XPath axes
> (and therefore the question which axes are redundant) in a research
> paper recently submitted for publication. Possibly this paper is of
> interest in the actual discussion which XPath axes shall be supported
> by the standards using XPath, or in XPath 2.0. We would be happy to
> hear comments from you on this paper. 
>  
> You can find the paper (PDF) at 
> 
> http://www.pms.informatik.uni-muenchen.de/publikationen/PMS-FB
> /PMS-FB-2001-16.pdf
> 
> and an abstract at the end of this mail.
> 
> Yours,
> Holger
> 
> 
> ====================================================
> 
> Abstract:
> 
> The location path language XPath is of particular importance for XML
> applications since it is a core component of many XML processing
> standards such as XSLT or XQuery.  In this paper, based on axis
> symmetry of XPath, equivalences of XPath 1.0 location paths involving
> reverse axes, such as ancestor and preceding, are established.
> These equivalences are used as rewriting rules in an algorithm for
> transforming location paths with reverse axes into equivalent
> reverse-axis-free ones.  Location paths without reverse axes as
> generated by the presented rewriting algorithm enable efficient
> SAX-like streamed data processing of XPath.
>  
>  
>

Received on Tuesday, 19 February 2002 07:52:24 UTC