the XMLeditor has been really interesting to work on lately. it seems to combine all of the approaches to parsing xml using an event parser, which alan and i have been exploring. Ultimately, we found trying to do several things at once became too complicated. So the model is now simplified to working one node at a time. Given our pseudo xpath, we can locate any node we want. We can return either the whole markup for a list view, or else just the fragment. We can take that fragment, build our xhtmlform structure and that is now editable, by simply replacing the old node with the new node. Delete is a case of replacing the old node with nothing. Insert is by far, the most complicated. It requires, first loading an empty node, which needs to be stored somewhere else (someday, it would be nice to try and construct that node from the schema…). Then upon insert, we need to fire user-defined functions to replace values for certain nodes. This helps to generate uids for the nodes. This is definitely possible. The approach i am trying is a way of registering a function for a given node (matched by our pseudo xpath). It’s almost like an event handler. interesting eh? i think i am beginning to see why java is interesting.
- Post published:April 14, 2004
- Post category:content management
Tags: content management, XML