XML has become the de facto standard format for data representation and exchange in domains ranging from the Web to desktop applications. Examples of XML-based document types include Geographic Information Systems Markup Language (GML) [8], Medical Markup Language (MML) [17], HL7 [10], and Open Document Format (ODF) [21]. This widespread use of XML requires efficient parsing techniques. The importance of efficient XML parsing methods was underscored by Nicola and John [19], showing that the parsing stage is processor and memory consuming, needing main memory as much as ten times the size of the original document. There are two de facto XML parsing APIs, DOM [2] and SAX [22]. SAX reads the whole document and generates a sequence of events according to the nesting of the elements, and hence it is not possible to skip reading parts of the document as this would change the semantics of the API. On the other hand, DOM allows users to explicitly navigate in the XML document using methods like