As I see it, there are 2 "main" options here: 1. Route the complete message (aka "Filtering") based on the result of an XPath evaluation. 2. Route a fragment of the message (or fragments, aka "splitting") based on an XPath expression. I think we definitely want #1. Perhaps people (Burr etc) also wanted #2 included as part of this. I think we should do #2 as a separate task. View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4256342#... Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&a...