Rik has some interesting observations regarding the depth of a present generation browser's "awareness" of the components that make up a Zope rendered html "page", which is to say nearly none. But isn't this exactly what XML solves for us? An XML document, even after browser rendering (not to be confused with the Zope server side composition stage), is still made up of discrete XML objects, and thus selectable, editable, etc. The beauty of this concept isn't so much that Zope / Python simply can't do all this by itself, we all know that Zope / Python (Grail?) can do anything short of levitation*. Rather the point is simply that XML is a well accepted, existing standard, and is bound to be supported in popular browsers. FWIW, Jerry S. * I missed the conference, so this is subject to updating ;-)