To say SAX is one-way and DOM is two-way is sort of true but really it is totally missing the point.
The 'truth' is that SAX is an event based parser, whereas DOM is a tree model parser.
SAX events could trigger output of a new document easily enough, in which case you could call it two way, just as DOM could be used purely to read and not to manipulate, in which case you could call it one way.
Your last post and the responses covered this ground though, I think you've got to realign your thoughts on these parsing models to what your previous responders and I have said -- that is SAX = event driven, DOM = tree model.