- From: Aaron Skonnard <aarons@develop.com>
- Date: Sun, 16 Apr 2000 00:47:55 -0600
- To: "www-Dom" <www-dom@w3.org>
I sent this to xml-dev to see if I'm the only one in the world that cares about this. ;-) -aaron BTW, I already sent the correction on the address (www-dom@w3.org) > -----Original Message----- > From: Aaron Skonnard [mailto:aarons@develop.com] > Sent: Saturday, April 15, 2000 5:34 PM > To: Xml-Dev@Xml. Org > Cc: www-dom-request@w3.org > Subject: adding addressing capabilities to the DOM > > > After looking over the recently released DOM Level 3 requirements > document, there has been some discussion on adding an addressing > API (similar to the selectNodes function in MSXML). The proposed > idea is to introduce a method on the Node interface that takes a > scheme specific addressing expression (e.g., XPath 1.0, the > product of the XQL WG, etc...) and returns a NodeList: > > NodeList getNodes(scheme, expression) > > The scheme identifies that language used in the expression. > Today, the main standard addressing language in place is XPath > 1.0, but this makes the API extensible for future versions of > XPath or other languages (similar to the way XPointer works). > With something like this in place, DOM developers can harness the > power of XPath to simplify typical navigation tasks in a standard way: > > NodeList books = doc.getNodes("xpath", > "/books//book[@topic='xml']/author"); > > The default scheme could also be specified for the entire > Document so it doesn't have to specified on each method call. > Similar functionality could also be added to the the Traversal > API (as another constructor type to NodeIterator and TreeWalker). > > In order to help the DOM WG decide on such an issue, the goal of > this email is to gather the general interest level from the xml-dev'ers. > > Who wants it? > > > BTW, if you have an opinion on this, please copy > www-dom-request@w3.org in your response. > > -aaron
Received on Sunday, 16 April 2000 02:49:22 UTC