Such an interface is useful if very large XML documents have to be processed and no DOM functions are required. By using this interface it is possible to read data stored within a XML document directly into the application datastructures without loading the document into memory.
The SAX interface of XML::LibXML is based on the famous XML::SAX interface. It uses the generic interface as provided by XML::SAX::Base.
Additionally to the generic functions, which are only able to process entire documents, XML::LibXML::SAX provides parse_chunk(). This method generates SAX events from well ballanced data such as is often provided by databases.
NOTE: At the moment XML::LibXML provides only an incomplete interface to libxml2's native SAX implementaion. The current implementation is not tested in production environment. It may causes significant memory problems or shows wrong behaviour. If you run into specific problems using this part of XML::LibXML, let me know.