Last modified: 14/10/2010
Unlike XSLT processors, OmniMark uses a streaming model to process XML/SGML.
A streaming model has advantages but it also has disadvantages. The XML flows from top to bottom. You can’t randomly access the XML tree in memory.
So how can you process an XML node in the beginning of a document if it depends on a node that hasn’t streamed yet or maybe never will come?