Nat Pryce wrote: > The DOM is a pretty awkward API to both use and implement. An API based on > Ruby-isms, such as dynamic typing and blocks for traversal, and XPath for > lookups would be much more convenient. we agree :) I talked about DOMs' data model, not the API. We're talking about an alternative to the DOM API (because it's awkward), currently dubbed "RubyWay tree API". It could look something like REXMLs' API, including XPath. This RubyWay tree API could build on the DOM data model (while possibly not using a single method of the DOM API), or it could (IMHO preferably) build on the Infoset data model. The Infoset is created for this task. Tobi -- Tobias Reif http://www.pinkjuice.com/myDigitalProfile.xhtml go_to('www.ruby-lang.org').get(ruby).play.create.have_fun http://www.pinkjuice.com/ruby/