If you only need to validate a document (and not do anything else with it) SAX does it without actually taking memory for the document: Just implement an empty handler which captures all SAX-events and ignores them. All but the error handling events.


All existing DOM parsers actually seem to be built on top of SAX parsers and a DOM is a very costly (but also valuable at times) data structure in memory. DOM comes in handy if you have to manipulate documents in memory.