All with appropriate parameters.
doc.html() view.html() doc.dxl() form.dxl() doc.rss() view.atom() doc.json() outline.json() etc...
along with Axel's XPath idea:
forall x in o.dxl(<xpath>)...
import/constructor methods would be nice too.
how often do you need the full yadayada when using the ?readviewentries?
it would be nice to be able to specify the option that you want the output without the timestamp, position, toplevelentries ... info