XML Lossy Text Compression: A Preliminary Study

Lossy compression techniques have been applied to image and text compression, yielding compression factors that are vastly superior to lossless compression schemes. In this paper, we present a preliminary study on a set of lossy transformations for XML documents that preserve the semantics. Inspired by previous techniques, e.g. lossy text compression and literate programming, we apply a simple algorithm to XML syntactic constructs to loose superfluous layout information and redundant text. The obtained XML keeps the human-readability and machine-readability properties. Additionally, it can lead to a considerable reduction of its space occupancy and boost the application of conventional text compressors, thus representing a promising technology for several data management tasks.