XPath Attribute Profile: metadata

["Cameron Morris" <cmorris@novell.com>]

We haven't explored meta-data publishing for xpath attributes and I think it will help clear some things up.

 

Robert -- yes, all I'm really interested in are the leaf nodes of the document.  And I think that is all we should try to accomplish with this profile, a way to map an XML document into one set of attributes.  

 

For meta-data I see two options:  

1) Enumerate in meta-data all xPath attributes(Leaf-nodes) that can be queried and asserted.

I like this because it does what I need it to do - and it's easy.  We don't have to restrict xPath with this route because there is only one set of allowable attributes.  This works with EP and PP because they have a rigid structure - no recursion.  Of cours, recursive structures are impossible to enumerate.  To solve this we can simply say no recursive documents.  

 

2) Enumerate in meta-data all fixed path nodes and flag attributes that allow additional xpath qualifiers at the end.

This handles any future recursive documents with the a suffix qualifier.  It also lets xPath aware SP's formulate specific xPaths above and beyond the enumerated ones specified by the AA.  

 

For example, Meta data for the an AA publishes

<attribute

   NameFormat="http://www.w3.org/TR/1999/REC-xpath-19991116"

   Name="/root/document"

   DocType="urn:rootdocument:schema"

   AllowSuffixQualifier="true" />

 

Then the SP could query for the following attribute:

 

<attribute

   NameFormat="http://www.w3.org/TR/1999/REC-xpath-19991116"

   Name="/root/document/*/@docID

   DocType="urn:rootdocument:schema"

/>

 

 

Both options remove the need for us (OASIS SSTC) to define an XPath restriction syntax.  It's up to the AA to publish what xpaths are supported.  (In the case of EP, PP and most documents the leaf nodes xpaths could be automatically generated).  It also limits the caching problem because the cache is limited to the enumerated xpaths.  If the suffix qualifier is on, then suffix qualified names will have the same problem that Anne Anderson mentioned.  Which is something I think we will have to live with.

 

- Cameron