XSLT without DOM tree

[Paulo Gaspar <pa...@krankikom.de>]

Remember M$ was talking about XSLT without DOM tree?


On new XSLT ideas and XML interfaces, just check the 
"XmlNavigator" idea on:
  http://msdn.microsoft.com/voices/xml07242000.asp

  <quote>
  XMLNavigator provides a "cursor" or "iterator" over 
  any XML tree, and methods for moving around the tree
  </quote>

Some methods:
  MoveToNext, MoveToPrevious, 
  MoveToParent, MoveToFirstChild, 
  MoveToAttribute, 
  Insert, Move, Remove, 
  CopyChildren,
  Select,
  Matches...

Another idea:
  <quote>
  XmlNavigator class can navigate any XML, not just the 
  in-memory DOM tree.
  </quote>

And from here... it gets really weird. Remember that M$
was thinking about implementing XSLT without DOM???
Well, that does NOT mean it is a SAX source either:
  <quote>
  ...by providing an XmlNavigator subclass called 
  DataDocumentNavigator, which can navigate relational 
  tables, rows, and columns stored in an ADO+ DataSet...
  </quote>
  <quote>
  The XmlNavigator class's XPath support also allows you 
  to provide input to the XslTransform class. Plug in 
  the DataDocumentNavigator, and you will quickly see 
  that you can now transform relational data with XSLT
  just as easily as you can transform the contents of an
  XmlDocument object.
  </quote>


And what about a XmlNavigator cursor parsing a large text 
file forward and backwards?
(Yeah... why not parsing backards? Even with some sparse
bookmarks or indexing helping a bit?)


Have fun,
Paulo Gaspar