That's only nicer if you have a use for XPath. That is one reason I wouldn't recommend rapidXML in use cases where the application is doing non-trivial edits of the DOM tree, because that's the kind of situation in which XPath is useful. But if your application consumes all the XML every time and writes the entire DOM tree every time, then XPath is of no utility. It's in some sense unfortunate, but the bottom line is that there isn't "the best" XML library, only more or less appropriate ones for particular use cases. I suspect this is a big contributor to the lack of XML support from Boost. At 05:38 AM 12/15/2008, Jose wrote:

Yes, rapidXML is quite nice. It was inspired by pugixml http://code.google.com/p/pugixml/, which is even nicer becaue it provides an xpath support. I personally find rapidXML not as useful/nice without xpath support. It's also quite easy to migrate from either one library to the other