> It's called an XML parser, Roman. It parses XML. The idea of using
> XML is so we don't have to create a parser for each new request type.

Do you have a universal parser that takes XML on one side and returns
my domain objects on the other? In general case this issue is not
solved in Java, did you manage that in C++?

Or you just have a SAX or DOM parser and application has either to
provide custom event SAX handlers for each XML structure or navigate
through the DOM to find the needed node (I doubt that you support
something like XPath there)?

Roman