Static Analysis for Event-Based XML Processing

Abstract

Event-based processing of XML data - as exemplified by the popular SAX framework - is a powerful alternative to using W3C's DOM or similar tree-based APIs. The event-based approach is particularly superior when processing large XML documents in a streaming fashion with minimal memory consumption.

This paper discusses challenges and presents some considerations for creating program analyses for SAX applications. In particular, we consider the problem of statically guaranteeing that a given SAX application always produces only well-formed and valid XML output.