Providing an Extensible Markup Language (XML) mapping framework to bridge the gap between heterogeneous sources of data is a challenging problem. The authors of this paper provide clear and motivating examples to show the complexity of XML mapping. An XML mapping architecture is presented.

The architecture is composed of four components. First, a user interface allows users to manage mapping visually, by providing various types of operations such as display, create, and modify. Second, there is an optional mapping population component, which is able to provide programmable mapping. Third, a mapping semantics component copes with mapping transformation semantics between the source and target models. Fourth, code generation supports the provision of executable and practical transformation code based on mapping semantics.

The paper is well written and structured, and describes a variety of current commercial mapping tools, though it suffers from a paucity of research-oriented references.