STnG — a Streaming Transformations and Glue framework

Abstract

STnG (pronounced “sting” ¹ ) is a framework for processing XML and other structured text. In developing STnG, it was our goal to allow complex transformations beyond those afforded by traditional XML transforming tools, such as XSLT, yet make the framework simple to use. We claim that to meet this goal, a system must:

1. support and encourage the use of small processing components
2. offer a hierarchical tree-like view of its data
3. factor out facilities for input chunking through a pattern/action model
4. not provide processing facilities of its own, instead invoking processors written in existing languages

STnG is built around common XML tools and idioms, but can process arbitrary structured text almost as easily as XML.

In the first part of this paper, we show how these requirements result in powerful and flexible systems, and how they can be achieved. The balance of this paper describes a processing framework we have developed in Java that implements these requirements. It is available for download at http://stng.sf.net.